# Claude Sonnet 4.6

> Source: https://aiwiki.ai/wiki/claude_sonnet_4_6
> Updated: 2026-10-28
> Fact-checked: 2026-11-02
> Categories: AI Models, Anthropic, Large Language Models
> License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) - attribute to "AI Wiki (aiwiki.ai)"
> Cite as: AI Wiki. "Claude Sonnet 4.6." aiwiki.ai, 28 Oct 2026. https://aiwiki.ai/wiki/claude_sonnet_4_6
> From AI Wiki (https://aiwiki.ai), the free encyclopedia of artificial intelligence. Reuse freely with attribution.

**Claude Sonnet 4.6** is a [large language model](https://aiwiki.ai/wiki/large_language_model) developed by [Anthropic](https://aiwiki.ai/wiki/anthropic) and released on February 17, 2026. It is the eighth model in the [Claude 4](https://aiwiki.ai/wiki/claude_4) generation and the immediate successor to [Claude Sonnet 4.5](https://aiwiki.ai/wiki/claude_sonnet_4_5), which had served as Anthropic's mid-tier flagship since September 2025. Sonnet 4.6 launched twelve days after [Claude Opus 4.6](https://aiwiki.ai/wiki/claude_opus_4_6) and brought the same one-million-token [context window](https://aiwiki.ai/wiki/context_window), [adaptive thinking](https://aiwiki.ai/wiki/adaptive_thinking), and context compaction features to the lower-priced Sonnet tier at unchanged pricing of $3 per million input tokens and $15 per million output tokens.[1][2]

<figure class="w-56">
<img src="/images/wd-2026-10-28-claude_2_1-logo-5276c950a6aa-tile.png" alt="Claude logo" width="480" height="156" />
<figcaption>Claude logo. Source: <a href="https://commons.wikimedia.org/wiki/File:Claude_AI_logo.svg">Wikimedia Commons</a>, <a href="https://commons.wikimedia.org/wiki/File:Claude_AI_logo.svg#Licensing">Public domain</a>.</figcaption>
</figure>

Anthropic positioned the release around a single claim, stating in the launch post that "performance that would have previously required reaching for an Opus-class model, including on real-world, economically valuable office tasks, is now available with Sonnet 4.6."[1] Databricks reported in the launch post that Sonnet 4.6 matches Opus 4.6 on OfficeQA, a benchmark that measures how well a model reads enterprise documents such as charts, PDFs, and tables and reasons from them.[1] The company reported that developers preferred Sonnet 4.6 to Sonnet 4.5 about 70% of the time in [Claude Code](https://aiwiki.ai/wiki/claude_code) testing and preferred it to [Claude Opus 4.5](https://aiwiki.ai/wiki/claude_opus_4_5) about 59% of the time on the same comparison.[1] On the [Artificial Analysis](https://aiwiki.ai/wiki/artificial_analysis) Intelligence Index, Sonnet 4.6 scored 51, two points behind Opus 4.6 and tied with [GPT-5.2](https://aiwiki.ai/wiki/gpt-5) (xhigh) at 51 points, narrowing what had been a seven-point gap between the Sonnet and Opus tiers in the prior generation to two points.[3]

The model became the default Claude on claude.ai and Claude Cowork for Free and Pro users on the day of release, replacing Sonnet 4.5 in that role. It also shipped on the same day to [Amazon Bedrock](https://aiwiki.ai/wiki/amazon_bedrock), [Google Cloud](https://aiwiki.ai/wiki/google_cloud) [Vertex AI](https://aiwiki.ai/wiki/google_vertex_ai), Microsoft Foundry, [GitHub Copilot](https://aiwiki.ai/wiki/github_copilot), Snowflake Cortex AI, and Anthropic's own Claude Code terminal agent, with day-one integrations from Cursor, [Replit](https://aiwiki.ai/wiki/replit), Hex, Factory, Notion, and several other developer platforms.[1][4][5][33]

The release was framed as a turning point for the Sonnet tier rather than a routine update. Artificial Analysis called Sonnet 4.6 the leader on [GDPval](https://aiwiki.ai/wiki/gdpval)-AA and TerminalBench, ahead of Opus 4.6 on those two evaluations despite the price gap.[3] DataCamp described it as approaching Opus-level intelligence at a price point that made it practical for far more tasks, while Caylent highlighted the jump in computer-use accuracy from 61.4% on Sonnet 4.5 to 72.5% on Sonnet 4.6 as the change with the largest practical implications for production agents.[6][7] [Zvi Mowshowitz](https://aiwiki.ai/wiki/zvi_mowshowitz) and others noted that real-world cost savings depended on effort tuning, since Sonnet 4.6 at max effort can over-spend output tokens.[8] On March 13, 2026, less than four weeks after launch, Anthropic moved the 1M context window to general availability and removed the long-context price surcharge entirely; a 900,000-token Sonnet 4.6 request was thereafter billed at the same per-token rate as a 9,000-token request.[21][22] By mid-May 2026, Sonnet 4.6 had become the standard Sonnet across nearly every major Claude distribution channel; [Microsoft 365 Copilot Studio](https://aiwiki.ai/wiki/microsoft_365_copilot) was scheduled to auto-migrate Sonnet 4.5 agents on May 30, 2026.[34] Anthropic succeeded it with [Claude Sonnet 5](https://aiwiki.ai/wiki/claude_sonnet_5) on June 30, 2026 and [Claude Sonnet 5.5](https://aiwiki.ai/wiki/claude_sonnet_5_5) on September 28, 2026; Sonnet 4.6 remains available as a legacy model.[23][54]

## Background

### The Sonnet product line

The Sonnet name within Anthropic's three-tier Claude family designates the mid-tier model: cheaper and faster than Opus, more capable than Haiku. The line began with [Claude 3 Sonnet](https://aiwiki.ai/wiki/claude_3_sonnet) in March 2024 and ran through [Claude 3.5 Sonnet](https://aiwiki.ai/wiki/claude_3_5_sonnet) (June 2024, refreshed October 2024), [Claude 3.7 Sonnet](https://aiwiki.ai/wiki/claude_3_7_sonnet) (February 2025), [Claude Sonnet 4](https://aiwiki.ai/wiki/claude_sonnet_4) (May 2025), and Claude Sonnet 4.5 (September 2025). Sonnet 4.6 is the sixth Sonnet release overall and the third within the Claude 4 generation. Sonnet pricing has held at $3 / $15 per million tokens since March 2024 despite substantial capability gains. Sonnet 4.5 was the first Sonnet to launch under [ASL-3](https://aiwiki.ai/wiki/responsible_scaling_policy) in September 2025, alongside context awareness, the agentic coding features that powered Claude Code 2.0, and Anthropic's [Claude Agent SDK](https://aiwiki.ai/wiki/claude_agent_sdk). Sonnet 4.6 inherits all those features and extends the line into 1M context windows and adaptive thinking, previously Opus-only.[1][9][10]

### Where does Sonnet 4.6 fit within Claude 4?

The Claude 4 family at Sonnet 4.6's release comprised seven publicly available models. [Claude Opus 4](https://aiwiki.ai/wiki/claude_opus_4) and Claude Sonnet 4 from May 2025 were still available; Anthropic deprecated them on April 14, 2026 and retired them on June 15, 2026. [Claude Opus 4.1](https://aiwiki.ai/wiki/claude_opus_4_1) (August 2025) and Claude Opus 4.5 (November 2025) remained available as legacy models. Claude Sonnet 4.5 (September 2025), [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5) (October 2025), and Claude Opus 4.6 (February 2026) were the active prior releases.[2][9] Sonnet 4.6 fits between Opus 4.6 and Haiku 4.5 in capability and price: 40% cheaper per token than Opus 4.6 with the same 1M context but a 64K output cap (versus 128K on Opus 4.6), trailing Opus 4.6 by one to two points on most benchmarks. Compared to Haiku 4.5 it is three times more expensive, with a five-times-larger context window and several percentage points stronger on reasoning, coding, and agentic benchmarks.[1][2][6]

The progression of recent Sonnet releases is summarized below.

| Sonnet model | Release date | Context window | Max output | Input ($/MTok) | Output ($/MTok) | SWE-bench Verified | OSWorld | ASL |
|---|---|---|---|---|---|---|---|---|
| [Claude 3.7 Sonnet](https://aiwiki.ai/wiki/claude_3_7_sonnet) | February 2025 | 200K | 64,000 | $3.00 | $15.00 | 70.3% | n/a | ASL-2 |
| [Claude Sonnet 4](https://aiwiki.ai/wiki/claude_sonnet_4) | May 22, 2025 | 200K | 64,000 | $3.00 | $15.00 | 72.7% | 42.2% | ASL-2 |
| [Claude Sonnet 4.5](https://aiwiki.ai/wiki/claude_sonnet_4_5) | September 29, 2025 | 200K | 64,000 | $3.00 | $15.00 | 77.2% | 61.4% | ASL-3 |
| **Claude Sonnet 4.6** | **February 17, 2026** | **1M** | **64,000** | **$3.00** | **$15.00** | **79.6%** | **72.5%** | **ASL-3** |

### Release context

Sonnet 4.6 arrived during a compressed window of Claude 4 family releases. Opus 4.5 had landed in November 2025 with the new $5 / $25 Opus pricing. Opus 4.6 followed on February 5, 2026, introducing adaptive thinking, context compaction, and the 1M-token context window. Sonnet 4.6 launched twelve days later, then long-context pricing went to standard rates on March 13, 2026, and [Claude Opus 4.7](https://aiwiki.ai/wiki/claude_opus_4_7) shipped on April 16, 2026 with a new tokenizer. With Sonnet 4.6 serving as the default in claude.ai (Free, Pro, Team) and in Claude Code for Pro and Team subscribers, it became the version of Claude that the largest share of users actually interact with, even after Opus 4.7 took over flagship benchmark headlines.[5][14][21][24] The April 9, 2026 GA of [Claude Cowork](https://aiwiki.ai/wiki/claude_cowork) on all paid plans extended Sonnet 4.6 into knowledge-work automation outside Claude Code, and the May 11, 2026 launch of [Claude Platform on AWS](https://aiwiki.ai/wiki/claude_platform_on_aws) added a third hyperscaler-billed surface.[35][36]

## What are the technical specifications of Claude Sonnet 4.6?

The table below summarizes the key parameters of Claude Sonnet 4.6 as documented in Anthropic's API documentation, the AWS Bedrock [model card](https://aiwiki.ai/wiki/model_card), and the system card.

| Parameter | Value |
|---|---|
| **Model ID ([Anthropic API](https://aiwiki.ai/wiki/anthropic_api))** | `claude-sonnet-4-6` |
| **Model alias** | `claude-sonnet-4-6` |
| **AWS Bedrock ID** | `anthropic.claude-sonnet-4-6` |
| **GCP Vertex AI ID** | `claude-sonnet-4-6` |
| **Release date** | February 17, 2026 |
| **[Context window](https://aiwiki.ai/wiki/context_window)** | 1,000,000 tokens (GA from March 13, 2026) |
| **Max output tokens** | 64,000 at launch; 128,000 per current API docs (300,000 via batch beta header) |
| **Input modalities** | Text, images |
| **Output modalities** | Text |
| **Extended thinking** | Yes |
| **Adaptive thinking** | Yes |
| **Computer use** | Supported |
| **Tool use / [function calling](https://aiwiki.ai/wiki/function_calling)** | Supported |
| **MCP support** | Supported |
| **Reliable knowledge cutoff** | August 2025 |
| **Training data cutoff** | January 2026 |
| **AI Safety Level** | ASL-3 |
| **[Prompt caching](https://aiwiki.ai/wiki/prompt_caching) (min tokens)** | 1,024 (Bedrock); 2,048 at launch on the [Anthropic API](https://aiwiki.ai/wiki/anthropic_api), 1,024 per current docs |
| **[Prompt caching](https://aiwiki.ai/wiki/prompt_caching) (max checkpoints)** | 4 per request |
| **[Prompt caching](https://aiwiki.ai/wiki/prompt_caching) cacheable fields (Bedrock)** | system, messages, tools |
| **Priority Tier** | Yes |
| **Latency tier** | Fast |

Unlike Opus 4.7, which shipped with a new tokenizer in April 2026 that increased per-task token usage by up to 35%, Sonnet 4.6 uses the same tokenizer as the rest of the Claude 4 generation. Per-token cost comparisons against Sonnet 4.5 therefore translate cleanly into per-task cost comparisons for the same workload.[1][2]

### Model ID format and snapshot

Starting with the Claude 4.6 generation, Anthropic moved away from snapshot-dated model identifiers in favor of a dateless format that is still a pinned snapshot rather than an evergreen pointer. The model ID `claude-sonnet-4-6` refers to a single fixed snapshot released on February 17, 2026, which Anthropic does not silently update. Future Sonnet releases will receive new dateless identifiers (for example, a hypothetical `claude-sonnet-4-7`) rather than continuing to update `claude-sonnet-4-6`.[2]

The Bedrock listing carries some additional notes. Bedrock supports Sonnet 4.6 through global cross-region inference (`global.anthropic.claude-sonnet-4-6`), four geographic-region IDs (`us`, `eu`, `au`, `jp`), and a London in-region option (`eu-west-2`). Geo cross-region inference is available from a wider region set including most US, EU, AU, and JP regions; all 30 advertised AWS regions worldwide can use the global inference profile. The Bedrock marketplace product ID is `prod-ffvjxvh4ltq64`. Standard and Reserved service tiers are supported; Priority and Flex are not at launch.[4]

### What is the one-million-token context window?

Sonnet 4.6 is the first Sonnet to support a one-million-token context window, matching Opus 4.6 and Opus 4.7. Anthropic describes this as enough to hold an entire codebase, a long contract, or dozens of research papers in a single request. At launch the feature was in beta on the [Anthropic API](https://aiwiki.ai/wiki/anthropic_api) and required usage tier 4 or above, with a long-context price surcharge for any input above 200,000 tokens.[2][11] On March 13, 2026, Anthropic moved the 1M context window to general availability for Sonnet 4.6 and Opus 4.6 and announced that standard pricing applies across the full window: a 900,000-token request is billed at the same per-token rate as a 9,000-token request. Full rate limits became available at every context length, [prompt caching](https://aiwiki.ai/wiki/prompt_caching) and batch processing discounts continued to apply at standard rates, and media capacity for image and PDF inputs expanded from 100 to 600 items per request.

### Adaptive thinking

Like Opus 4.6, Sonnet 4.6 uses adaptive thinking instead of the fixed `budget_tokens` model that Sonnet 4.5 inherited from Claude 3.7 Sonnet. The model decides at runtime whether and how deeply to reason; developers choose an effort level (low, medium, high, max) instead of a hard token budget. Low effort suits chat and low-latency agents; high is the API default, though Anthropic recommends medium for most applications; high allocates more tokens for complex multi-step problems; max permits tens of thousands of reasoning tokens on a single response with corresponding cost.[1][12] Artificial Analysis found that at max effort Sonnet 4.6 used about 28% more output tokens than Opus 4.6 across its Intelligence Index, partly explaining why the headline 40% per-token discount versus Opus 4.6 narrowed in real evaluation runs.[3] Sonnet 4.5 deployments that passed `thinking={"type": "enabled", "budget_tokens": 32000}` continue to work for backward compatibility; new integrations should use `thinking={"type": "adaptive"}` plus an `effort` setting.[12]

### Context compaction

Context compaction, introduced for Opus 4.6 and inherited here, lets long-running agents summarize earlier turns automatically once they approach the context-window limit. With compaction enabled, an agent can run for hundreds of turns or thousands of tool calls before hitting an explicit context overflow. Default thresholds are 50,000 tokens minimum and 150,000 tokens default trigger, both configurable.[1][2][5] Alexlavaee.me described compaction as allowing "effectively infinite conversations."[5]

### Vision and computer use

Sonnet 4.6 supports image input alongside text, with the same vision pipeline as the rest of the Claude 4 generation: document analysis, chart and figure understanding, screenshot interpretation, and [visual question answering](https://aiwiki.ai/wiki/visual_question_answering). Vision integrates tightly with computer use, where the model reads desktop or browser screenshots and produces keyboard and mouse actions for multi-step UI tasks.[1][6] On OSWorld-Verified, Sonnet 4.6 scored 72.5%, narrowly behind Opus 4.6 (72.7%), well ahead of Sonnet 4.5 (61.4%), and dramatically above the original Sonnet computer-use baseline of around 14.9%. Insurance-automation company Pace reported 94% accuracy on its own insurance computer-use benchmark.[1][3][5][7]

### Tool use, function calling, and MCP

Sonnet 4.6 supports structured tool use, parallel tool calls, the [Model Context Protocol](https://aiwiki.ai/wiki/model_context_protocol), and interleaved reasoning between tool calls. As of April 2026 the pricing documentation listed per-request token overhead of 346 tokens for the [system prompt](https://aiwiki.ai/wiki/system_prompt) when `tool_choice` is `auto` or `none` and 313 tokens when `tool_choice` is `any` or a specific tool; the current docs list 497 and 589 tokens. Server-side tools add their own overheads: 245 tokens for the bash tool, 700 tokens for the text editor (`text_editor_20250429`), 735 tokens per computer-use tool definition (with an additional 466 to 499 tokens of system prompt). These figures matter when budgeting prompt-cache hit rates and per-call token costs.[1][2] Developers using Claude Agent SDK or third-party agent frameworks can switch from Sonnet 4.5 to Sonnet 4.6 by changing only the model identifier; the API surface is backward compatible within the Sonnet line.[2][12]

## How much does Claude Sonnet 4.6 cost?

Claude Sonnet 4.6 uses standard per-token billing with no subscription requirement for API access. Pricing is unchanged from Sonnet 4.5 and remained at the standard rate after long-context pricing went GA on March 13, 2026.

| Tier | Input ($/MTok) | Output ($/MTok) |
|---|---|---|
| **Standard (any context length, GA from March 13, 2026)** | $3.00 | $15.00 |
| **Long context (>200K input, prior to March 13, 2026)** | $6.00 | $22.50 |
| **Batch API (50% discount)** | $1.50 | $7.50 |
| **Prompt cache writes (5-minute TTL, 1.25x base)** | $3.75 | n/a |
| **Prompt cache writes (1-hour TTL, 2x base)** | $6.00 | n/a |
| **Prompt cache reads (0.1x base)** | $0.30 | n/a |
| **US data residency multiplier ([Anthropic API](https://aiwiki.ai/wiki/anthropic_api))** | 1.1x | 1.1x |
| **AWS Bedrock regional/multi-region premium** | 1.1x | 1.1x |

Anthropic markets up to 90% cost savings via prompt caching and 50% via batch processing. The cache-read rate of $0.30/MTok is one tenth of the standard input rate, breaking even after one cached read for the 5-minute TTL or two reads for the 1-hour TTL. Two pricing modifiers stack: US-only inference via `inference_geo` (1.1x across all categories) and Bedrock/Vertex AI regional or multi-region endpoints (1.1x versus global), introduced with Sonnet 4.5 in late 2025.[2][11] Most Bedrock and Vertex AI listings match Anthropic API rates closely; [OpenRouter](https://aiwiki.ai/wiki/openrouter) quotes the standard $3 / $15.

### Cost relative to sibling Claude models

The table below compares Sonnet 4.6 to the rest of the Claude 4 generation on a per-token basis at the post-March-13 standard rate.

| Model | Input ($/MTok) | Output ($/MTok) | Long-context premium | Ratio vs Sonnet 4.6 input | Ratio vs Sonnet 4.6 output |
|---|---|---|---|---|---|
| [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5) | $1.00 | $5.00 | n/a | 0.33x | 0.33x |
| **Claude Sonnet 4.6** | **$3.00** | **$15.00** | **None (standard pricing across full 1M window)** | **1.0x** | **1.0x** |
| [Claude Sonnet 4.5](https://aiwiki.ai/wiki/claude_sonnet_4_5) | $3.00 | $15.00 | n/a (200K only) | 1.0x | 1.0x |
| [Claude Opus 4.5](https://aiwiki.ai/wiki/claude_opus_4_5) | $5.00 | $25.00 | n/a | 1.67x | 1.67x |
| [Claude Opus 4.6](https://aiwiki.ai/wiki/claude_opus_4_6) | $5.00 | $25.00 | None (standard pricing across full 1M window) | 1.67x | 1.67x |
| [Claude Opus 4.7](https://aiwiki.ai/wiki/claude_opus_4_7) | $5.00 | $25.00 | None (standard pricing across full 1M window) | 1.67x | 1.67x |

At list price, Sonnet 4.6 is 40% cheaper per token than Opus and 200% more expensive than Haiku. Real-world differences vary depending on effort levels and long context use. Artificial Analysis reported Sonnet 4.6 used about 74 million output tokens to complete its full Intelligence Index suite, roughly 28% more than Opus 4.6's 58M and three times Sonnet 4.5's 25M, partly offsetting per-token savings. All-in cost was $2,088 for Sonnet 4.6, $733 for Sonnet 4.5, and $2,486 for Opus 4.6: Sonnet 4.6 was only 16% cheaper than Opus 4.6 at max effort on this benchmark, a much smaller discount than the headline 40% per-token gap. For workloads dominated by short tasks where deep reasoning rarely fires, the per-token gap translates more directly.[3]

The April 9, 2026 advisor strategy offered another cost-quality trade-off: Sonnet 4.6 or Haiku 4.5 runs a task as executor and consults Opus as an advisor when stuck. In Anthropic's evaluations, with Opus 4.6 as advisor, Sonnet 4.6 scored 2.7 points higher on SWE-bench Multilingual than Sonnet alone while cutting cost per agentic task by 11.9%.[53]

### Code execution, web search, and managed agent pricing

Beyond messaging, several built-in tools have separate fees. Code execution is free when invoked with web search or web fetch; otherwise execution time is billed against an organization-wide allowance of 1,550 free hours per month, with overage at $0.05 per container-hour. Web search is $10 per 1,000 search calls plus standard token costs for retrieved content. Claude Managed Agents add a session runtime fee of $0.08 per running session-hour on top of standard token rates and do not get the Batch API discount or the Fast mode premium.[2]

## How does Claude Sonnet 4.6 perform on benchmarks?

The table below presents the most widely cited benchmark scores for Claude Sonnet 4.6 alongside Sonnet 4.5 (its predecessor) and Opus 4.6 (the contemporary flagship). Where a benchmark was reported in the Anthropic launch post, the headline figure is used; where it was reported by an independent benchmark aggregator, the source is cited.

| Benchmark | Sonnet 4.6 | Sonnet 4.5 | Opus 4.6 | Source |
|---|---|---|---|---|
| **[SWE-bench Verified](https://aiwiki.ai/wiki/swe-bench_verified)** | 79.6% | 77.2% | 80.8% | [Anthropic](https://aiwiki.ai/wiki/anthropic) launch post[1] |
| **[SWE-bench Multilingual](https://aiwiki.ai/wiki/swe_bench_multilingual)** | 75.9% | n/a | n/a | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **OSWorld-Verified (computer use)** | 72.5% | 61.4% | 72.7% | [Anthropic](https://aiwiki.ai/wiki/anthropic) launch post[1] |
| **Terminal-Bench 2.0** | 59.1% | 51.0% | 65.4% | DataCamp / Caylent[6][7] |
| **[GPQA Diamond](https://aiwiki.ai/wiki/gpqa_diamond)** | 89.9% | 83.4% | 91.3% | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **AIME 2025 (no tools)** | 95.6% | n/a | n/a | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **ARC-AGI-1** | 86% | n/a | n/a | Mowshowitz analysis[8] |
| **[ARC-AGI-2](https://aiwiki.ai/wiki/arc_agi_2)** | 60.4% (or 58%) | n/a | 68.8% | NxCode / Mowshowitz[8][13] |
| **GDPval-AA (Elo)** | 1633 | n/a | 1606 | [Artificial Analysis](https://aiwiki.ai/wiki/artificial_analysis)[3] |
| **TerminalBench ([Artificial Analysis](https://aiwiki.ai/wiki/artificial_analysis))** | 53% | n/a | 46% | Artificial Analysis[3] |
| **Insurance computer use (Pace benchmark)** | 94% | n/a | n/a | Pace, quoted in launch post[1] |
| **Finance Agent v1.1** | 63.3% | n/a | 60.1% | DataCamp[6] |
| **MCP-Atlas (tool use)** | 61.3% | 43.8% | 59.5% (max effort) | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **tau2-bench Retail** | 91.7% | 86.2% | 91.9% | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **tau2-bench Telecom** | 97.9% | 98.0% | 99.3% | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card[20] |
| **BenchLM TAU-bench aggregate (April 2026 snapshot)** | 87.5% | 86.2% | 84.8% | BenchLM[38] |
| **[BrowseComp](https://aiwiki.ai/wiki/browsecomp) (single-agent)** | 74.0% | n/a | 84.0% | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card update[27] |
| **[BrowseComp](https://aiwiki.ai/wiki/browsecomp) (multi-agent)** | 82.07% | n/a | n/a | [Anthropic](https://aiwiki.ai/wiki/anthropic) system card update[27] |
| **SRE-Skills-Bench** | 90.4% | 85.9% | 94.7% | Rootly[28] |
| **NYT Connections** | 58% | 49% | substantially higher | Mowshowitz analysis[8] |
| **Artificial Analysis Intelligence Index** | 51 | 43 | 53 | Artificial Analysis[3] |
| **Single-turn harmlessness** | 99.38% | 97.89% | similar | System card[20] |
| **[Prompt injection](https://aiwiki.ai/wiki/prompt_injection) attack success (no safeguards)** | 1.29% | 49.36% | similar | System card[20] |

Notes: [SWE-bench Verified](https://aiwiki.ai/wiki/swe-bench_verified) scores are the headline single-run figures from Anthropic's launch post. The OSWorld-Verified score for Sonnet 4.6 is the harmonized public number; OSWorld-Verified is a curated subset of the full OSWorld benchmark. [ARC-AGI-2](https://aiwiki.ai/wiki/arc_agi_2) figures vary slightly between sources (60.4% from NxCode, 58% from Mowshowitz's read of the same release), because Anthropic's chart shows the max-effort score (58.3%) and its footnote reports 60.4% at high effort. Terminal-Bench 2.0 figures (59.1% for Sonnet 4.6) match Anthropic's system card table and are repeated by DataCamp and Caylent. Anthropic released a revised system card on March 6, 2026 after running an improved cheating-detection pipeline that found nine additional unintended solutions in the original [BrowseComp](https://aiwiki.ai/wiki/browsecomp) evaluation, leading the single-agent score to be revised from 74.72% down to 74.01% and the multi-agent score from 82.62% down to 82.07%; the table above uses the revised figures.[1][3][6][7][8][13][27]

### SWE-bench Verified, Multilingual, and Pro

[SWE-bench](https://aiwiki.ai/wiki/swe_bench) Verified tests real-world [GitHub](https://aiwiki.ai/wiki/github) issue resolution where the model must identify a bug or feature gap and write a patch that passes the repository's test suite. Sonnet 4.6's 79.6% (the widely aggregated figure) is a 2.4-point improvement over Sonnet 4.5 (77.2%) and lands within 1.2 points of Opus 4.6 (80.8%). Anthropic's own launch post reports its primary score averaged over 10 trials and notes that "with a prompt modification, we saw a score of 80.2%."[1] It is the highest SWE-bench Verified score reported for any Sonnet-class model up to that point.[1][14] Caylent described the gap as near-parity, and DataCamp noted that recent frontier models all cluster around 80% on the benchmark.[6][7] Boris Cherny, the Claude Code creator, was reported to still prefer Opus for all coding work, reasoning that human time spent correcting mistakes matters more than token cost.[5]

On the harder [SWE-bench Pro](https://aiwiki.ai/wiki/swe_bench_pro) split, Anthropic did not report a Sonnet 4.6 score, and the Morphllm aggregate lists none; the 56.8% figure sometimes attributed to Sonnet 4.6 is GPT-5.3 Codex's self-reported score. The same aggregate placed Opus 4.7 at 64.3% and [Gemini 3.1 Pro](https://aiwiki.ai/wiki/gemini_3_1_pro) at 54.2%. [SWE-bench Multilingual](https://aiwiki.ai/wiki/swe_bench_multilingual) from the Anthropic system card gave Sonnet 4.6 a 75.9% pass rate, a few points below the Verified split.[20][37]

The post-launch landscape placed Opus 4.6 at 80.8%, Gemini 3.1 Pro at 80.6%, MiniMax M2.5 at 80.2%, Sonnet 4.6 at 79.6%, and [Grok 4](https://aiwiki.ai/wiki/grok_4) at 73.5% on SWE-bench Verified. The cluster between 79% and 81% indicated that leading agentic-coding models had converged at the Verified level, with differentiation moving to the harder SWE-bench Pro split and recent agentic terminal benchmarks.[26][29]

### OSWorld-Verified and computer use

OSWorld-Verified measures navigation of desktop and web GUIs, form filling, and multi-step computer tasks. Sonnet 4.6's 72.5% is the largest single capability jump in the launch: Sonnet 4.5 had scored 61.4% five months earlier, and Sonnet 4 scored only 42.2% in May 2025. Sonnet 4.6 is essentially tied with Opus 4.6 (72.7%) and well ahead of GPT-5.3 Codex (64.7%). Pace, an insurance-automation company, reported in the launch post that Sonnet 4.6 hit 94% on its insurance computer-use benchmark, the highest of any model it had tested.[1][3][6][7]

### GDPval-AA and Terminal-Bench

GDPval-AA is Artificial Analysis's Elo-style evaluation of economically valuable knowledge-work tasks (drafting documents, building spreadsheets, analyst work). Sonnet 4.6 scored 1633 Elo, ahead of Opus 4.6 at 1606 and every other model AA tested at launch. The result was the strongest evidence that Sonnet 4.6 had reached or exceeded Opus-class capability on the practical office tasks that drive [enterprise AI](https://aiwiki.ai/wiki/enterprise_ai) usage.[3]

On Terminal-Bench 2.0, Sonnet 4.6 scored 59.1%, an 8.1-point improvement over Sonnet 4.5's 51.0%. Opus 4.6 leads at 65.4%, and GPT-5.3 Codex was reported at 75.1%, the only headline-tier evaluation where the Opus and Sonnet advantage over a non-Anthropic model was clearly reversed.[6][7][29] Artificial Analysis's parallel TerminalBench evaluation gave Sonnet 4.6 53%, ahead of Opus 4.6 at 46%, likely reflecting differences in scoring or subset.[3]

### Tau-Bench, tau2-bench, and MCP-Atlas tool use

[Tau-Bench](https://aiwiki.ai/wiki/tau-bench) from Sierra tests realistic customer-service workflows by calling tools, reading results, and reasoning across many turns. On tau2-bench, Anthropic's system card reports 91.7% on retail and 97.9% on telecom for Sonnet 4.6, versus 91.9% and 99.3% for Opus 4.6 and 86.2% and 98.0% for Sonnet 4.5.[20] BenchLM's April 10, 2026 snapshot of an aggregate TAU-bench score listed Sonnet 4.6 at 87.5%, behind the invitation-only [Claude Mythos Preview](https://aiwiki.ai/wiki/claude_mythos) at 89.2% and ahead of Sonnet 4.5's 86.2%; BenchLM later withdrew that table for lacking exact source attachments.[38] On MCP-Atlas, Sonnet 4.6 scored 61.3%, ahead of Opus 4.6's 59.5% at max effort, although Opus 4.6 reached 62.7% at high effort.[20]

### BrowseComp and web research

BrowseComp is the live-web information-seeking benchmark with 1,266 tasks requiring navigation of the live internet. Anthropic's revised system card from March 6, 2026 reported Sonnet 4.6 at 74.01% single-agent and 82.07% multi-agent, after an improved cheating-detection pipeline flagged nine additional unintended solutions (revising the single-agent score from 74.72%) and 11 in the multi-agent run (revising it from 82.62%). The results place Sonnet 4.6 behind Opus 4.6 (84.0%) and [GPT-5.2](https://aiwiki.ai/wiki/gpt-5.2) (77.9%) single-agent but ahead of most of the field in multi-agent configurations where multiple instances coordinate to verify answers. 

### GPQA, Math, MMLU-Pro, HumanEval, and ARC-AGI

Graduate-level reasoning is one area where Sonnet 4.6 still trails Opus 4.6, though narrowly. On [GPQA](https://aiwiki.ai/wiki/gpqa) Diamond, Anthropic's system card reports 89.9% for Sonnet 4.6, up from Sonnet 4.5's 83.4% and slightly below Opus 4.6's 91.3%; the same table lists GPT-5.2 at 93.2% and Gemini 3 Pro at 91.9%, and Gemini 3.1 Pro later reported 94.3%.[20][30] (NxCode's guide lists a much lower 74.1% for Sonnet 4.6.[13]) On mathematics, Anthropic's system card reports 95.6% on AIME 2025 without tools, while warning that contamination may have inflated the score.[20]

On [ARC-AGI](https://aiwiki.ai/wiki/arc_agi), Sonnet 4.6 scored 86% on ARC-AGI-1 and either 58% (Mowshowitz) or 60.4% (NxCode) on ARC-AGI-2, with the difference attributable to evaluation harness or effort settings. ARC Prize's Greg Kamradt, quoted by Mowshowitz, said Sonnet 4.6 delivered "less performance than Opus 4.6 (expected), but for around the same cost (unexpected)," with per-task ARC-AGI-1 at about $1.45 and ARC-AGI-2 at $2.72.[8]

### Domain benchmarks: Finance, SRE, Insurance

On Finance Agent v1.1, Sonnet 4.6 scored 63.3%, leading the field and beating Opus 4.6 at 60.1%.[6] On SRE-Skills-Bench from Rootly, Sonnet 4.6 scored 90.4% versus Sonnet 4.5's 85.9% and Opus 4.6's 94.7%. The per-domain breakdown was instructive: Sonnet 4.6 led on general SRE (88.0% vs 87.0%), tied on AWS networking (97.1%), and trailed on AWS S3 (75.7% vs 91.9%) and AWS IAM (85.2% vs 92.2%). On root-cause accuracy in production incident reproductions, Rootly reported Sonnet 4.6 performing similarly to Opus 4.6 and beating it in some cases despite running at 40% lower cost per token.[28] Pace reported that Sonnet 4.6 scored 94% on its insurance computer-use benchmark, the highest-performing model it had tested for computer use.[1]

### Other reasoning and creative benchmarks

On NYT Connections, a creative-reasoning probe, Sonnet 4.6 scored 58%, up from Sonnet 4.5's 49% but well below Opus 4.6.[8] On the Artificial Analysis Intelligence Index, Sonnet 4.6 reached 51 points (Sonnet 4.5: 43; Opus 4.6: 53), tying GPT-5.2 (xhigh). Artificial Analysis noted that the Sonnet-Opus gap had narrowed from seven points (Sonnet 4.5 vs Opus 4.5) to two, and that Anthropic held the top two spots on its index for the first time.[3]

### Speed and throughput

One developer quoted by Mowshowitz reported that with thinking disabled, Sonnet 4.6's time to first token was significantly faster and lower-variance than Sonnet 4.5's, on par with Haiku 4.5.[8] Anthropic positions Sonnet 4.6 as Fast latency tier, between the Fastest Haiku 4.5 and Moderate Opus 4.6/4.7. Opus 4.6 has a separate Fast mode beta at 6x standard pricing; no equivalent fast mode tier is documented for Sonnet 4.6.[2][6]

## How does Claude Sonnet 4.6 compare with other Claude models?

The table below compares Sonnet 4.6 to current and legacy Claude models on the dimensions most relevant to deployment decisions.

| Model | Input ($/MTok) | Output ($/MTok) | Context window | Max output | Extended thinking | Adaptive thinking | Computer use | Latency tier |
|---|---|---|---|---|---|---|---|---|
| [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5) | $1.00 | $5.00 | 200K | 64K | Yes | No | Yes | Fastest |
| [Claude Sonnet 4.5](https://aiwiki.ai/wiki/claude_sonnet_4_5) | $3.00 | $15.00 | 200K | 64K | Yes | No | Yes | Fast |
| **Claude Sonnet 4.6** | **$3.00** | **$15.00** | **1M** | **64K** | **Yes** | **Yes** | **Yes** | **Fast** |
| [Claude Opus 4.5](https://aiwiki.ai/wiki/claude_opus_4_5) | $5.00 | $25.00 | 200K | 64K | Yes | No | Yes | Moderate |
| [Claude Opus 4.6](https://aiwiki.ai/wiki/claude_opus_4_6) | $5.00 | $25.00 | 1M | 128K | Yes | Yes | Yes | Moderate |
| [Claude Opus 4.7](https://aiwiki.ai/wiki/claude_opus_4_7) | $5.00 | $25.00 | 1M | 128K | No (adaptive only) | Yes | Yes | Moderate |

Compared to Sonnet 4.5, Sonnet 4.6 has the same pricing and the same maximum output token cap but a 5x larger context window, plus adaptive thinking and context compaction. The two models are otherwise drop-in compatible at the API level; an application calling Sonnet 4.5 typically needs only the model identifier change and the optional addition of an effort level to migrate.[2][12]

Compared to Opus 4.6, Sonnet 4.6 is 40% cheaper per token, has the same context window but a smaller output cap (64K vs 128K), and trails Opus 4.6 by roughly one to two percentage points on most benchmarks except GDPval-AA, MCP-Atlas, Finance Agent v1.1, and TerminalBench (Artificial Analysis), where Sonnet 4.6 leads. For agentic tasks, Artificial Analysis found Sonnet 4.6 the better price-performance choice over Opus 4.6; for deep scientific reasoning, GPQA, ARC-AGI-2, BrowseComp single-agent, and Terminal-Bench 2.0 in absolute terms, Opus 4.6 retains its advantage.[3][6][7]

Compared to Opus 4.7, Sonnet 4.6 is similarly 40% cheaper but trails by a wider margin on coding benchmarks. Opus 4.7's SWE-bench Verified of 87.6% is eight percentage points above Sonnet 4.6. Opus 4.7 also has higher OSWorld accuracy (78.0% versus 72.5%) and a 1:1 vision pixel mapping that Sonnet 4.6 does not have. Anthropic positions Opus 4.7 as the model of choice for the most complex agentic coding work and Sonnet 4.6 as the right default for everyday workloads, including most of Claude Code's day-to-day usage.[14][15][37]

### Comparison with non-Anthropic frontier models

By Q2 2026, Sonnet 4.6 was being benchmarked against GPT-5 variants and [Gemini](https://aiwiki.ai/wiki/gemini) 3 Pro/3.1 Pro on most of the same evaluations. Sonnet 4.6 led on agentic computer use (OSWorld 72.5% vs GPT-5.3 Codex 64.7%) and agentic office tasks (GDPval-AA Elo 1633 vs next-best 1606), led on several tool-use benchmarks (MCP-Atlas 61.3%, tau2-bench retail 91.7%) while trailing GPT-5.2 slightly on tau2-bench telecom (97.9% vs 98.7%), and trailed on graduate-level scientific reasoning (Gemini 3.1 Pro reached 94.3% on [GPQA Diamond](https://aiwiki.ai/wiki/gpqa_diamond)) and Terminal-Bench 2.0 (GPT-5.3 Codex 75.1%).[7][13][29][30][38]

BenchLM's comparison against [GPT-5.5](https://aiwiki.ai/wiki/gpt-5.5) shows GPT-5.5 ahead in most categories.[39]

Gemini 3.1 Pro's list price ($12 per million output tokens, per the Morphllm price table) undercuts Sonnet 4.6's $15 output rate.[37]

## Capabilities in detail

### Agentic coding

Sonnet 4.6's 79.6% on SWE-bench Verified places it at near-flagship coding accuracy. Anthropic's launch post highlighted the model's behavior across long agentic coding sessions: it follows multi-step plans more reliably, makes fewer false claims of completion, and is less prone to over-engineering simple changes than its predecessor.[1] DataCamp's hands-on review ran three small tests (a multi-step number puzzle, a matrix rotation and a Python debugging task); Sonnet 4.6 passed all three, performing on par with Opus 4.6 in those examples.[6]

Claude Code shipped Sonnet 4.6 support on launch day, with the model picker allowing users to choose between Sonnet 4.6 and Opus 4.6.[1][16] GitHub Copilot made Sonnet 4.6 generally available across Pro, Pro+, Business, and Enterprise plans on the same date, with support spanning [Visual Studio Code](https://aiwiki.ai/wiki/visual_studio_code) (chat, ask, edit, agent), Visual Studio (agent, ask), github.com, GitHub Mobile, the Copilot CLI, and the Copilot [coding agent](https://aiwiki.ai/wiki/coding_agent). The integration launched at a 1x premium-request multiplier described as tentative.[16]

GitHub's Joe Binder, VP of Product, said that "out of the gate, Claude Sonnet 4.6 is already excelling at complex code fixes, especially when searching across large codebases is essential," a claim consistent with the strong long-context retrieval and near-Opus SWE-bench score.[1][16] Replit president Michele Catasta described Sonnet 4.6 as having an "extraordinary" performance-to-cost ratio and noted that "it's hard to overstate how fast Claude models have been evolving in recent months."[1]

### Computer use

The move from 61.4% to 72.5% on OSWorld-Verified is the single largest capability change in the Sonnet 4.6 release in proportional terms. Pace reported 94% accuracy on its insurance computer-use benchmark, and Opus 4.6 reaches similar OSWorld accuracy at 67% higher cost per token. Alexlavaee.me reported a roughly 5x improvement in OSWorld accuracy over the original computer-use baseline from October 2024 (around 14.9%).[1][3][5]

As with Opus 4.6, Anthropic recommends that production computer-use deployments include human-in-the-loop oversight for high-stakes actions (financial transactions, file deletions, irreversible operations) and explicit retry, validation, and rollback layers for low-stakes tasks. The recommended pattern is "approval workflows + retries + validation" rather than fully autonomous operation.[7]

### Long-horizon agents and planning

Sonnet 4.6 inherits the long-horizon agent design that Sonnet 4.5 introduced in September 2025 and the agent-team architecture that Opus 4.6 introduced in February 2026. With context compaction enabled, a Sonnet 4.6 agent maintains coherent state across hundreds of turns or thousands of tool calls without an external memory layer. With adaptive thinking, it scales reasoning depth dynamically rather than spending the same effort on every sub-step. Anthropic's launch post illustrated long-horizon planning with Vending-Bench Arena, a simulated business competition in which Sonnet 4.6 invested heavily in capacity for the first ten simulated months and then pivoted to profitability, finishing well ahead of the competition.[1]

In the multi-agent configuration that Anthropic reported on BrowseComp, multiple Sonnet 4.6 instances coordinated to research and verify answers. The 82.07% multi-agent score beat the best single-agent configuration (74.01%) by 8.1 percentage points.[27]

A second multi-model pattern emerged on April 9, 2026 with the advisor strategy: a Sonnet 4.6 executor consults Opus (Opus 4.6 in Anthropic's tests) only when it needs guidance, which Anthropic said raised SWE-bench Multilingual by 2.7 points and cut cost per agentic task by 11.9% compared with Sonnet alone.[53]

### Vision and document understanding

Sonnet 4.6 processes image inputs alongside text including documents, charts, screenshots, and photos. The model carries forward the same vision pipeline as Sonnet 4.5 and Opus 4.6, distinct from the new tokenizer and high-resolution vision arriving with Opus 4.7.[15]  The March 13 GA announcement raised per-request media capacity from 100 to 600 images or PDF pages.[21]

### Multilingual capability

Claude Sonnet 4.6 supports the same major world languages as the Claude 4 family: English, Spanish, French, German, Portuguese, Italian, Dutch, Japanese, Korean, Chinese. Anthropic's system card reports 89.3% on MMMLU (multilingual [MMLU](https://aiwiki.ai/wiki/mmlu)), averaged over 10 trials across 14 non-English languages, compared with 89.5% for Sonnet 4.5 and 91.1% for Opus 4.6.[20] On SWE-bench Multilingual specifically, Sonnet 4.6 scored 75.9%, a few points below its 79.6% on the Verified split and consistent with the regression most frontier coding models show when moving outside Python.[20]

### Prompt caching and batch processing

Sonnet 4.6 supports prompt caching with a minimum cacheable prompt of 2,048 tokens on the Anthropic API at launch (1,024 per current docs) or 1,024 tokens on Bedrock, and up to four cache checkpoints per request. TTL is five minutes or one hour. Cache reads bill at one-tenth the input rate, so a 200,000-token system prompt costing $0.75 to write at the 5-minute TTL is $0.06 per cached read.[2][4][11] Cacheable fields on Bedrock include system prompt, messages, and tools blocks; the Anthropic API adds explicit cache breakpoints on individual content blocks. Batch processing through the Message Batches API gives a 50% discount with up to 24-hour turnaround. Sonnet 4.6 supports up to 300,000 output tokens per batched request via the `output-300k-2026-03-24` beta header. Caching and batch discounts stack with each other but not with managed-agent runtime billing.[2]

## Use cases

### Default model on claude.ai

On February 17, 2026, Sonnet 4.6 replaced Sonnet 4.5 as the default Claude on claude.ai for Free and Pro users. Pro and Max users continued to have access to Opus 4.6 as a premium model selectable from the picker. Anthropic also upgraded the free tier to include file creation, connectors, skills and compaction.[1]

### Daily coding driver

Claude Code shipped Sonnet 4.6 as the default for most users on launch day, with Opus 4.6 selectable for harder problems. Hex Technologies CTO Caitlin Colgrove said the company was moving the majority of its traffic to Sonnet 4.6, seeing Opus-level performance on all but its hardest analytical tasks, and Factory said it was transitioning its Sonnet traffic to the new model. Pasquale Pillitteri's Claude Code guide recommended Sonnet 4.6 as the default, switching to Opus only when absolute maximum performance is needed. Cursor's documentation took a similar position: Sonnet 4.6 "is better on longer tasks but below Opus for raw intelligence."[1][5][18][32]

### Document and analyst workflows

Sonnet 4.6's strength on GDPval-AA and on the insurance computer-use benchmark made it a natural fit for document-processing and analyst workflows in regulated industries: contract review, financial analyst tasks, policy comparison.[6] After the March 13 1M context GA, the cost calculus for very long workflows changed materially. For example, 700,000 input tokens that would have cost $4.20 at the prior $6 long-context input rate cost $2.10 at the standard $3 rate, and output on such requests dropped from $22.50 to $15 per million tokens. The same reduction applies to large-codebase security audits, multi-paper research synthesis, and similar long-form analyst tasks.[21][22]

### Multi-agent orchestration

The Sonnet 4.5-plans, Haiku 4.5-executes pattern from October 2025 carries forward with Sonnet 4.6 capable of taking the orchestrator role for non-flagship workloads where Opus 4.6 would be overkill. The larger context window lets a Sonnet 4.6 orchestrator maintain coherent global state for longer runs without external memory.[1][7] The BrowseComp multi-agent result (82.07%) demonstrated the technique inside the same tier.[27] In the April 2026 advisor-strategy variant, a Sonnet 4.6 executor selectively consults Opus, and Anthropic reported an 11.9% lower cost per agentic task than Sonnet alone.[53]

### AI SRE and incident response

Rootly's evaluation cited Sonnet 4.6's 90.4% on SRE-Skills-Bench (up from 85.9%) and near-parity with Opus 4.6 on production incident root-cause investigations as evidence that the model had become a viable default for AI SRE products at the same per-token cost. The 1M context lets long incident logs and infrastructure manifests fit in a single request. Per-domain breakdown still favors Opus 4.6 on AWS S3 (91.9% vs 75.7%) and AWS IAM (92.2% vs 85.2%), suggesting infrastructure investigations crossing those services may benefit from escalation.[28]

### Microsoft 365 and Snowflake enterprise channels

Microsoft 365 Copilot added Sonnet 4.6 alongside its [OpenAI](https://aiwiki.ai/wiki/openai) defaults on March 9, 2026, which one blogger framed as giving roughly 400 million Office users access to Anthropic's model. Microsoft 365 Copilot Studio retired Sonnet 4.5 on May 30, 2026 with auto-migration to Sonnet 4.6 (one-time 30-day deferral through June 30 available).[34][41][42] On May 1, 2026, Microsoft 365 E7, "the Frontier Suite", reached general availability as a bundle of Microsoft 365 Copilot, Microsoft 365 E5, Microsoft Entra Suite and [Microsoft Agent 365](https://aiwiki.ai/wiki/microsoft_agent_365).[51] When [Microsoft](https://aiwiki.ai/wiki/microsoft) announced E7 on March 9, 2026, it said Claude was available in Copilot's mainline chat through its Frontier program and that Copilot Cowork, built with Anthropic, was in research preview.[52]

Snowflake Cortex AI shipped Sonnet 4.6 on launch day under a multi-year $200 million partnership announced in December 2025, making Claude available to Snowflake's 12,600+ customers; Sonnet 4.6 launched in private preview in the U.S. for Cortex Code, Cortex AI Functions and the Cortex REST API, with Snowflake Intelligence support to follow.[33][43]

## Reception and adoption

Reception was strongly positive on technical merits, with coverage focused on three threads: the closing gap between the Sonnet and Opus tiers, the leap in computer-use accuracy, and price stability. DataCamp described Sonnet 4.6 as approaching Opus-level intelligence at a price point that made it practical for far more tasks.[6] Caylent wrote that at about 72% first-attempt success, "approval workflows + retries + validation layers turn computer-use into reliable throughput," while cautioning that it is still not capable of running unattended in production.[7] Artificial Analysis called it "the overall strongest model we have tested for agentic use cases" while noting its higher token use.[3] NxCode framed it as "79.6% SWE-bench at $3/MTok."[13]

Developer testimonials in the launch post and follow-up coverage included Hex Technologies CTO Caitlin Colgrove ("We're moving the majority of our traffic to Sonnet 4.6"), Factory.AI (transitioning primary agent traffic), Replit's Michele Catasta ("extraordinary" performance-to-cost ratio), GitHub's Joe Binder ("excelling at complex code fixes, especially when searching across large codebases is essential"), and Hercules CEO Brendan Falk ("Opus 4.6 level accuracy, instruction following, and UI, all for a meaningfully lower cost"). One early-access design customer said Sonnet 4.6 "has perfect design taste when building frontend pages and data reports, and it requires far less hand-holding to get there than anything we've tested before."[1][5][16]

Mowshowitz provided the most detailed independent analysis, headlining his post "Claude Sonnet 4.6 Gives You Flexibility." Mowshowitz wrote that Sonnet 4.6 is "modestly cheaper and faster than Opus 4.6, and for most purposes it's modestly not as good," and said he would generally stick with Opus while using Sonnet for computer use or high-volume work where it is good enough. Reactions he collected included one user calling the model "overfitted on agenticity" and another finding it "very (over) eager to call tools," and he noted Sonnet can waste tokens at max effort, as on ARC. He also cited Anthropic's Drake Thomas on lower sycophancy and a user who found it far more reluctant to help craft prompt injections.[8]

Not all reactions were uniformly enthusiastic. Boris Cherny, the Claude Code creator, was reported as preferring Opus for his own use, arguing the SWE-bench gap mattered for solo debugging.[5] In reactions collected by Mowshowitz, some users reported Sonnet 4.6 being over-eager to call tools, including when they were not needed.[8]

A mid-May 2026 Winbuzzer report flagged that Microsoft was reducing internal Claude Code licenses for some engineering teams in favor of GitHub Copilot CLI, framed as internal tooling consolidation. Microsoft 365 Copilot's Claude integration, the Copilot Studio Sonnet 4.6 default, and the GitHub Copilot Sonnet 4.6 listing remained unaffected.[44]

### Adoption and market context

By late April 2026, multiple data points pointed to broad uptake. Hex and Factory publicly disclosed moving traffic to the model. The March 13 1M-context GA announcement paired standard pricing with full rate limits at every context length, signalling that production capacity for long-context workloads had matured to the point where Anthropic was comfortable removing previously tier-gated access.[1][5][16][21][22][32]

Aggregate adoption numbers from third-party trackers gave further context. Demandsage reported the Claude API processed more than 25 billion calls per month by Q1 2026, 45% from enterprise platforms, and Anthropic's annualized run-rate revenue reached roughly $14 billion in February 2026. The Claude API served over 87,000 organizations by Q1 2026, a 142% year-over-year increase, with Anthropic counting 300,000+ business customers.[45][46][47] [Salesforce](https://aiwiki.ai/wiki/salesforce)'s April 2026 Developer Edition refresh added Agentforce Vibes IDE with Claude Sonnet 4.5 (not Sonnet 4.6) as the default coding model.[48]

## Is Claude Sonnet 4.6 safe?

Anthropic deployed Claude Sonnet 4.6 under AI Safety Level 3 (ASL-3), the same standard applied to Sonnet 4.5, Opus 4.5, Opus 4.6, and Opus 4.7. Anthropic's launch communication described Sonnet 4.6 as having "a broadly warm, honest, prosocial, and at times funny character, very strong safety behaviors, and no signs of major concerns around high-stakes forms of misalignment."[1] Anthropic describes its ASL-3 deployment standard as primarily about risks from chemical and biological weapons from threat actors with relatively modest resources and expertise. The classification is conservative: Anthropic emphasizes that ASL-3 is applied when the company cannot rule out the relevant capability threshold rather than as a definitive determination that the model has crossed it.[1][19]

### Single-turn harmlessness

The system card reports 99.38% harmless-response rate on single-turn violative requests, up from Sonnet 4.5's 97.89% and broadly comparable to Opus 4.6. On the harder violative subset the rate is 99.40% (vs 98.40% for Sonnet 4.5). Over-refusal on benign prompts is 0.18% (higher-difficulty) and 0.41% (straightforward), versus Sonnet 4.5's 8.50% and 0.08%; the 8.50% figure had been one of the most-cited concerns about Sonnet 4.5 and the new number addresses it directly. On agentic coding tests with malicious intent, Sonnet 4.6 achieved a 100% refusal rate. The child-safety subset shows 99.96% harmless on violative prompts and 0.08% refusal on benign prompts.[7][20]

### Prompt injection

Resistance to prompt injection saw the largest absolute change of any safety metric. On Anthropic's standard browser-environment prompt-injection benchmark, Sonnet 4.6 had an attack success rate of 1.29% of scenarios (0.29% of attempts) without external safeguards, versus Sonnet 4.5's 49.36% of scenarios (16.23% of attempts). With safeguards enabled, the success rate fell to 0.51% of scenarios and 0.08% of attempts. Anthropic said Sonnet 4.6 "performs similarly to Opus 4.6" on prompt-injection resistance, and the system card called it "comparable to Claude Opus 4.6" with the new safeguards enabled.[1][20] The reduction from a roughly 50% scenario-level success rate on Sonnet 4.5 to a roughly 1% rate brings Sonnet-tier deployments much closer to a security profile compatible with sensitive enterprise data, though Anthropic still recommends production deployments stack input and output classifier safeguards on top.[7][20]

### Misalignment evaluations

The system card reports Sonnet 4.6 showed lower rates of misaligned behavior than Sonnet 4.5 on Anthropic's automated alignment evaluations, including the [agentic misalignment](https://aiwiki.ai/wiki/agentic_misalignment) scenarios that drew attention with the Opus 4 system card in May 2025. On sycophancy, researcher Drake Thomas noted Sonnet 4.6 was less sycophantic than all prior Claude models including Opus 4.6, with "warmth" modestly higher than Sonnet 4.5 but lower than Opus 4.6.[8][20]

The system card reports "elevated but not extreme" verbalized evaluation awareness (an average of 1.2 on a 1-10 scale), and the reactions Mowshowitz collected pointed to lower sycophancy and stronger [prompt injection](https://aiwiki.ai/wiki/prompt_injection) resistance. Anthropic's launch communication called the character "broadly warm, honest, prosocial, and at times funny," with "very strong safety behaviors, and no signs of major concerns around high-stakes forms of misalignment."[1][8][20]

### CBRN, cybersecurity, and autonomy

The system card includes evaluations for biological, chemical, radiological, and nuclear uplift, cybersecurity, and autonomous capability. Anthropic deployed Sonnet 4.6 under the ASL-3 Standard. On AI R&D-4, it said it did not believe the model met the threshold but was "in a gray zone where clean rule-out is difficult and the margin to the threshold is unclear." Deployment-time safeguards include the same input and output classifiers rolled out for Opus 4.6, plus the constitutional-AI-derived training that the rest of the Claude 4 family shares. For cybersecurity, Sonnet 4.6 is competent on entry-level cyber tasks but does not reliably solve advanced offensive scenarios; the highest-capability cybersecurity-relevant model is held back as the invitation-only [Project Glasswing](https://aiwiki.ai/wiki/project_glasswing) Mythos Preview, which crossed 93.9% on SWE-bench Verified internally but remains restricted to defensive cybersecurity uplift.[15][19][20]

### System card revision

On March 6, 2026, Anthropic published a revised system card after running an improved cheating-detection pipeline on BrowseComp. The new pipeline identified nine additional unintended shortcuts and revised the single-agent BrowseComp score from 74.72% to 74.01% and the multi-agent score from 82.62% to 82.07%, after 11 flagged multi-agent problems were re-run with the leaks removed.[27]

## Where is Claude Sonnet 4.6 available?

The deployment matrix below summarizes current options as of May 2026.

| Channel | Model identifier | Notes |
|---|---|---|
| [Anthropic API](https://aiwiki.ai/wiki/anthropic_api) (1P) | `claude-sonnet-4-6` | Messages API, Batches API, [prompt caching](https://aiwiki.ai/wiki/prompt_caching), MCP, all betas |
| Claude Managed Agents | `claude-sonnet-4-6` | Adds $0.08/hour session runtime; Batch and Fast mode unavailable |
| Claude Platform on AWS | `claude-sonnet-4-6` ([Anthropic](https://aiwiki.ai/wiki/anthropic) console, AWS auth) | Launched May 11, 2026; most AWS commercial regions; AWS billing and IAM |
| AWS Bedrock (global) | `global.anthropic.claude-sonnet-4-6` | Routes worldwide; available in 30+ regions |
| AWS Bedrock (geo) | `us / eu / au / jp.anthropic.claude-sonnet-4-6` | Single geography for data residency; 1.1x premium |
| AWS Bedrock (in-region) | `anthropic.claude-sonnet-4-6` in eu-west-2 | London in-region for strict residency; 1.1x premium |
| GCP Vertex AI | `claude-sonnet-4-6` | Global, multi-region, regional endpoints; 1.1x premium non-global |
| Snowflake Cortex AI | Anthropic Claude Sonnet 4.6 | Day-one; $200M expanded partnership; 12,600+ enterprise customers |
| Microsoft Foundry | Claude Sonnet 4.6 in catalog | Available from launch |
| [Microsoft 365 Copilot](https://aiwiki.ai/wiki/microsoft_365_copilot) | Sonnet 4.6 in model picker | Live March 9, 2026; ~400M Office seats |
| Microsoft 365 Copilot Studio | Sonnet 4.6 (default after May 30, 2026) | Sonnet 4.5 retires May 30; auto-migration; 30-day deferral option |
| claude.ai | default on Free and Pro | Free tier includes file creation, connectors, skills, compaction |
| [Claude Code](https://aiwiki.ai/wiki/claude_code) | default for Pro/Team | Opus 4.6 selectable; from March 13, Opus 4.6 sessions get 1M context on Max, Team and Enterprise |
| [Claude Cowork](https://aiwiki.ai/wiki/claude_cowork) | default underlying model | GA April 9, 2026 on all paid plans; role-based access, Zoom MCP, group spend |
| [GitHub Copilot](https://aiwiki.ai/wiki/github_copilot) | generally available across plans | VS Code, Visual Studio, GitHub Mobile, Copilot CLI, [coding agent](https://aiwiki.ai/wiki/coding_agent) |
| Cursor | `claude-4-6-sonnet` | Default for most agent-mode coding sessions |
| Third-party | [Replit](https://aiwiki.ai/wiki/replit), Hex, Factory.AI, Notion, [Augment Code](https://aiwiki.ai/wiki/augment_code), Warp, Zencoder, Poe | Day-one or near-day-one integrations |

The Anthropic API exposes the most beta features (300K-token batch outputs, prompt cache breakpoints, [MCP server](https://aiwiki.ai/wiki/mcp_server)s, latest computer-use beta) but requires Anthropic's own billing. AWS Bedrock and Vertex AI offer the same model with platform-native billing, data residency at a 1.1x premium, and cloud IAM integration. [Claude Platform on AWS](https://aiwiki.ai/wiki/claude_platform_on_aws), generally available since May 11, 2026, gives AWS customers Anthropic's native Claude platform (Console, Managed Agents, Skills, Files API, MCP connector, advisor strategy) using AWS authentication and billing, with availability planned in most AWS commercial regions. The choice is typically a procurement decision rather than a capability one.[2][4][36]

## Snapshots and version history

| Snapshot | API model ID | Release date | Status (May 2026) |
|---|---|---|---|
| Initial release | `claude-sonnet-4-6` | February 17, 2026 | Active |

No second Sonnet 4.6 snapshot was released. Anthropic released [Claude Sonnet 5](https://aiwiki.ai/wiki/claude_sonnet_5) on June 30, 2026, describing it as a substantial improvement over Sonnet 4.6, and [Claude Sonnet 5.5](https://aiwiki.ai/wiki/claude_sonnet_5_5) on September 28, 2026; Sonnet 4.6 remains available on the API as a legacy model, with retirement scheduled no sooner than February 17, 2027.[23][54] A source map file accidentally shipped in the Claude Code npm package on March 31, 2026 referenced a `claude-sonnet-4-8` identifier, leading several outlets to report that Anthropic plans to skip a hypothetical Sonnet 4.7 numbering; no Sonnet 4.8 was released, and the next Sonnet model was Claude Sonnet 5 on June 30, 2026.[23][49][50] The February 17, 2026 snapshot remains the only Sonnet-tier model in the Claude 4 generation beyond Sonnet 4.5. The dateless identifier convention introduced with the 4.6 generation means future updates will use a new dateless ID rather than reusing `claude-sonnet-4-6`. The system card was revised on March 6, 2026; the underlying model snapshot was not changed.[2][27]

### Timeline of Sonnet 4.6 events

| Date | Event |
|---|---|
| February 5, 2026 | [Claude Opus 4.6](https://aiwiki.ai/wiki/claude_opus_4_6) released; introduces adaptive thinking, context compaction, 1M context beta |
| February 17, 2026 | Claude Sonnet 4.6 released; default on claude.ai for Free, Pro, and Team |
| February 17, 2026 | [GitHub Copilot](https://aiwiki.ai/wiki/github_copilot), Cursor, [Replit](https://aiwiki.ai/wiki/replit), Hex, Factory, Notion, AWS Bedrock, GCP Vertex AI, Microsoft Foundry, Snowflake Cortex AI day-one launches |
| March 6, 2026 | System card revised after improved [BrowseComp](https://aiwiki.ai/wiki/browsecomp) cheating-detection pipeline; scores adjusted |
| March 9, 2026 | [Microsoft 365 Copilot](https://aiwiki.ai/wiki/microsoft_365_copilot) integration live; Sonnet 4.6 reaches ~400M Office seats |
| March 13, 2026 | 1M [context window](https://aiwiki.ai/wiki/context_window) moves to GA at standard pricing for Sonnet 4.6 and Opus 4.6; long-context surcharge removed; image/PDF capacity raised to 600 per request |
| March 31, 2026 | Source map leak references `claude-sonnet-4-8` identifier inside [Claude Code](https://aiwiki.ai/wiki/claude_code) npm package (unconfirmed by [Anthropic](https://aiwiki.ai/wiki/anthropic)) |
| April 9, 2026 | [Claude Cowork](https://aiwiki.ai/wiki/claude_cowork) reaches general availability on all paid plans; advisor tool introduced, with Sonnet 4.6 as executor and Opus 4.6 as advisor in Anthropic's tests |
| April 16, 2026 | [Claude Opus 4.7](https://aiwiki.ai/wiki/claude_opus_4_7) released; new tokenizer; Sonnet 4.6 remains active default Sonnet |
| April 30, 2026 | Microsoft Copilot Studio Sonnet 4.5 evaluation window opens; admins can test agents on Sonnet 4.6 |
| May 1, 2026 | Microsoft 365 E7 Frontier Suite GA, bundling [Microsoft 365 Copilot](https://aiwiki.ai/wiki/microsoft_365_copilot), E5, Entra Suite and Agent 365 |
| May 11, 2026 | Claude Platform on AWS launches with Sonnet 4.6 |
| May 30, 2026 | Scheduled retirement of Sonnet 4.5 in Microsoft Copilot Studio; auto-migration to Sonnet 4.6 begins |
| May 2026 | Sonnet 4.6 remains the active Sonnet flagship |
| June 30, 2026 | [Claude Sonnet 5](https://aiwiki.ai/wiki/claude_sonnet_5) released, succeeding Sonnet 4.6 as the current Sonnet |
| September 28, 2026 | [Claude Sonnet 5.5](https://aiwiki.ai/wiki/claude_sonnet_5_5) released; Sonnet 4.6 listed as a legacy model, retirement not sooner than February 17, 2027 |

## Migration guide

Migrating from Sonnet 4.5 to Sonnet 4.6 typically requires only changing the model identifier. Two areas warrant attention. First, code that set `thinking={"type": "enabled", "budget_tokens": <n>}` on Sonnet 4.5 should migrate to adaptive thinking via `thinking={"type": "adaptive"}` plus an `effort` hint (low, medium, high, max), starting with `medium` and raising to `high` for tasks that previously needed deep extended thinking.[12] Second, Sonnet 4.6 at higher effort levels can spend significantly more reasoning tokens than Sonnet 4.5 did under the prior fixed-budget regime. Artificial Analysis's all-in cost data (Sonnet 4.6 at $2,088 versus Sonnet 4.5 at $733 on the same Intelligence Index suite) is the most-cited reference; production cost monitoring should track per-task output-token usage during the migration window.[3][12]

Prompt caching, batch API, and MCP-based tool-use code requires no changes; the same multipliers apply. For Bedrock, the identifier change from `anthropic.claude-sonnet-4-5-20250929-v1:0` to `anthropic.claude-sonnet-4-6` plus a region or endpoint selection is sufficient. Code that previously used a regional endpoint should review whether the global endpoint with 10% lower pricing is acceptable for its data residency requirements.[2][4][12]

Microsoft Copilot Studio tenants face a separate cutover: Sonnet 4.5 retires on May 30, 2026 with auto-migration to Sonnet 4.6 unless administrators enable the one-time 30-day deferral through June 30. The April 30 to May 30 evaluation window is the recommended testing slot.[34][42]

## Limitations

### Smaller output cap than Opus 4.6

At launch Claude Sonnet 4.6 had a 64,000-token maximum output, half of Opus 4.6's 128,000 tokens; Anthropic's current model page lists 128,000 tokens for Sonnet 4.6.[54] For long-form generation use cases such as book-length manuscripts, very long contracts, or large codebase rewrites, the output cap can require chunking or multi-call orchestration even when the input fits in a single 1M-token request. The Batch API beta header `output-300k-2026-03-24` partially addresses this for non-real-time workloads, allowing up to 300,000 output tokens per batched request.[2]

### High-effort cost

Multiple independent reviews found that Sonnet 4.6 at max effort can spend output tokens at a rate that erodes the 40% per-token discount versus Opus 4.6. Artificial Analysis reported that Sonnet 4.6 used roughly 28% more output tokens than Opus 4.6 to complete the same Intelligence Index suite, and Mowshowitz found ARC-AGI-2 evaluation costs roughly equal between Sonnet at max effort and Opus. Production deployments that use max effort routinely should benchmark real-world cost rather than relying on per-token list price.[3][8]

### GPQA Diamond figures

NxCode's benchmark guide listed Sonnet 4.6 at 74.1% on GPQA Diamond, but Anthropic's system card reports 89.9% (averaged over 10 trials, max effort), above Sonnet 4.5's 83.4% and 1.4 points below Opus 4.6's 91.3%. Third-party collations of this benchmark should be checked against the system card.[13][20]

### Computer-use accuracy still requires oversight

While 72.5% on OSWorld-Verified is a major leap, it still implies that roughly 27% of complex GUI automation tasks fail or require intervention. Anthropic recommends production computer-use deployments include human-in-the-loop oversight for high-stakes actions and explicit retry-and-validation layers for low-stakes tasks. Caylent's review described the design pattern as "approval workflows + retries + validation," not autonomous operation.[1][7]

### Tool-calling eagerness

Mowshowitz and several developer-forum commenters reported that Sonnet 4.6 sometimes called tools more aggressively than Sonnet 4.5 in agent loops, occasionally invoking tools that were not necessary for the immediate task. The behavior is broadly positive in environments designed around heavy tool use but can show up as unnecessary cost or latency in lighter-weight applications.[8]

### Knowledge cutoff

The model's reliable knowledge cutoff is August 2025 and its training data cutoff is January 2026. Applications requiring up-to-date information about events after January 2026 should supplement the model with [retrieval-augmented generation](https://aiwiki.ai/wiki/retrieval_augmented_generation) or live tool access to current sources.[2]

### No high-resolution vision

Claude Opus 4.7, released two months after Sonnet 4.6, introduced a new tokenizer and high-resolution vision pipeline (up to 2,576 pixels on the long edge with 1:1 pixel-to-coordinate mapping). Sonnet 4.6 retains the older Claude 4 tokenizer and vision pipeline, with vision capped at 1,568 pixels on the long edge. Use cases requiring fine-grained vision, such as clicking on small UI elements or reading low-resolution scanned text, may need to use Opus 4.7 instead.[15]

### No fast mode tier

Opus 4.6 has a Fast mode beta available at 6x standard pricing for cases where developers prefer to pay for higher tokens-per-second at flagship intelligence. Sonnet 4.6 does not have an equivalent fast mode tier as of May 2026; users requiring sub-Fast latency at Sonnet quality must rely on Anthropic's Priority Tier or on the standard latency profile.[2]

### Widening coding gap to Opus 4.7

The arrival of Opus 4.7 on April 16, 2026 widened the headline coding gap between Sonnet and Opus inside the Claude 4 family: eight points on SWE-bench Verified (87.6% vs 79.6%). For workloads where Sonnet 4.6's coding accuracy is the binding constraint, the advisor strategy is an intermediate option, but applications requiring Opus-level coding accuracy on every step should price in the full Opus 4.7 cost.[15][35][37]

## See also

- [Anthropic](https://aiwiki.ai/wiki/anthropic)
- [Claude](https://aiwiki.ai/wiki/claude)
- [Claude 4](https://aiwiki.ai/wiki/claude_4)
- [Claude Sonnet 4.5](https://aiwiki.ai/wiki/claude_sonnet_4_5)
- [Claude Sonnet 4](https://aiwiki.ai/wiki/claude_sonnet_4)
- [Claude Opus 4.6](https://aiwiki.ai/wiki/claude_opus_4_6)
- [Claude Opus 4.7](https://aiwiki.ai/wiki/claude_opus_4_7)
- [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5)
- [Claude Code](https://aiwiki.ai/wiki/claude_code)
- [Claude Cowork](https://aiwiki.ai/wiki/claude_cowork)
- [Claude Platform on AWS](https://aiwiki.ai/wiki/claude_platform_on_aws)
- [Model Context Protocol](https://aiwiki.ai/wiki/model_context_protocol)
- [Responsible Scaling Policy](https://aiwiki.ai/wiki/responsible_scaling_policy)
- [SWE-bench](https://aiwiki.ai/wiki/swe-bench)
- [Tau-Bench](https://aiwiki.ai/wiki/tau-bench)
- [GPQA](https://aiwiki.ai/wiki/gpqa)
- [GPT-5](https://aiwiki.ai/wiki/gpt-5)
- [Gemini](https://aiwiki.ai/wiki/gemini)
- [Gemini 3](https://aiwiki.ai/wiki/gemini_3)
- [Microsoft 365 Copilot](https://aiwiki.ai/wiki/microsoft_365_copilot)
- [Project Glasswing](https://aiwiki.ai/wiki/project_glasswing)

## References

1. Anthropic. "Introducing Claude Sonnet 4.6." Anthropic Newsroom, February 17, 2026. https://www.anthropic.com/news/claude-sonnet-4-6
2. Anthropic. "Models overview." Claude API Documentation. Retrieved May 2026. https://platform.claude.com/docs/en/about-claude/models/overview
3. Artificial Analysis. "Claude Sonnet 4.6: Everything You Need to Know." February 2026. https://artificialanalysis.ai/articles/sonnet-4-6-everything-you-need-to-know
4. Amazon Web Services. "Claude Sonnet 4.6 model card." Amazon Bedrock Documentation. https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-sonnet-4-6.html
5. Lavaee, Alex. "Claude Sonnet 4.6: What Developers Actually Need to Know." alexlavaee.me, February 2026. https://alexlavaee.me/blog/sonnet-4-6-technical-breakdown/
6. DataCamp. "Claude Sonnet 4.6: Features, Access, Tests, and Benchmarks." February 2026. https://www.datacamp.com/blog/claude-sonnet-4-6
7. Caylent. "Claude Sonnet 4.6 in Production: Capability, Safety, and Cost Explained." February 2026. https://caylent.com/blog/claude-sonnet-4-6-in-production-capability-safety-and-cost-explained
8. Mowshowitz, Zvi. "Claude Sonnet 4.6 Gives You Flexibility." Don't Worry About the Vase (Substack), February 2026. https://thezvi.substack.com/p/claude-sonnet-46-gives-you-flexibility
9. Wikipedia contributors. "Claude (language model)." Wikipedia, retrieved May 2026. https://en.wikipedia.org/wiki/Claude_(language_model)
10. Anthropic. "Introducing Claude Sonnet 4.5." Anthropic Newsroom, September 29, 2025. https://www.anthropic.com/news/claude-sonnet-4-5
11. Anthropic. "Claude Sonnet 4.6 product page." Anthropic. Retrieved May 2026. https://www.anthropic.com/claude/sonnet
12. Anthropic. "Adaptive thinking." Claude API Documentation. https://platform.claude.com/docs/en/build-with-claude/adaptive-thinking
13. NxCode. "Claude Sonnet 4.6: 79.6% SWE-bench at $3/MTok, Complete Guide (2026)." 2026. https://www.nxcode.io/resources/news/claude-sonnet-4-6-complete-guide-benchmarks-pricing-2026
14. Anthropic. "Introducing Claude Opus 4.6." Anthropic Newsroom, February 5, 2026. https://www.anthropic.com/news/claude-opus-4-6
15. Anthropic. "Introducing Claude Opus 4.7." Anthropic Newsroom, April 16, 2026. https://www.anthropic.com/news/claude-opus-4-7
16. GitHub. "Claude Sonnet 4.6 is now generally available in GitHub Copilot." GitHub Changelog, February 17, 2026. https://github.blog/changelog/2026-02-17-claude-sonnet-4-6-is-now-generally-available-in-github-copilot/
17. Anthropic. "Claude Sonnet 4.6 product page." Anthropic. Retrieved May 2026. https://www.anthropic.com/claude/sonnet
18. Pillitteri, Pasquale. "How to Use Claude Sonnet 4.6 in Claude Code: Complete Guide." 2026. https://pasqualepillitteri.it/en/news/292/claude-sonnet-4-6-claude-code-guide
19. Anthropic. "Responsible Scaling Policy Version 3.0." Anthropic Newsroom, 2025. https://www.anthropic.com/news/responsible-scaling-policy-v3
20. Anthropic. "System Card: Claude Sonnet 4.6." February 17, 2026. https://www-cdn.anthropic.com/bbd8ef16d70b7a1665f14f306ee88b53f686aa75.pdf
21. Anthropic. "1M context window now generally available at standard pricing." Anthropic blog, March 13, 2026. https://claude.com/blog/1m-context-ga
22. The New Stack. "Anthropic makes a pricing change that matters for Claude's longest prompts." March 2026. https://thenewstack.io/claude-million-token-pricing/
23. Anthropic. "Release notes." Claude Help Center. Retrieved May 2026. https://support.claude.com/en/articles/12138966-release-notes
24. FetchLogic. "Claude Sonnet 4.6: Anthropic's 2026 Power Move Explained." February 2026. https://fetchlogic.net/claude-sonnet-4-6-is-anthropics-quiet-power-move-and-free-users-are-the-trojan-horse/
25. PricePerToken. "HumanEval Leaderboard 2026 - Compare AI Model Scores." Retrieved May 2026. https://pricepertoken.com/leaderboards/benchmark/humaneval
26. Morphllm. "Claude Benchmarks (2026): Every Score for Opus 4.6, Sonnet 4.6 & Haiku." 2026. https://www.morphllm.com/claude-benchmarks
27. Anthropic. "Claude Sonnet 4.6 system card revision (BrowseComp cheating-detection update)." March 6, 2026. https://www.anthropic.com/claude-sonnet-4-6-system-card
28. Rootly. "Claude Sonnet 4.6: Benchmark Results and Lessons for AI SRE." February 18, 2026. https://rootly.com/blog/claude-sonnet-4-6-benchmark-results-and-lessons-for-ai-sre
29. Marc0.dev. "SWE-Bench Verified Leaderboard March 2026." Retrieved May 2026. https://www.marc0.dev/en/leaderboard
30. AceCloud. "Gemini 3.1 Pro Vs Sonnet 4.6 Vs Opus 4.6 Vs GPT-5.2 Vs Meta Muse Spark (2026)." 2026. https://acecloud.ai/blog/gemini-3-1-pro-vs-sonnet-4-6-vs-opus-4-6-vs-gpt-5-2/
31. IntuitionLabs. "Claude vs ChatGPT vs Copilot vs Gemini: 2026 Enterprise Guide." 2026. https://intuitionlabs.ai/articles/claude-vs-chatgpt-vs-copilot-vs-gemini-enterprise-comparison
32. Cursor. "Claude 4.6 Sonnet documentation." Cursor Docs. Retrieved May 2026. https://cursor.com/docs/models/claude-4-6-sonnet
33. Snowflake. "Announcing Anthropic Claude Sonnet 4.6 on Snowflake Cortex AI." Snowflake blog, February 2026. https://www.snowflake.com/en/blog/claude-sonnet-4-6-snowflake-cortex-ai/
34. Microsoft 365 Message Center. "Microsoft 365 Copilot: Claude Sonnet 4.5 retiring for Copilot Studio (automatic migration to Claude Sonnet 4.6)" (MC1296875), April 2026. https://www.itelio.com/en/microsoft-message-center/MC1296875
35. Pillitteri, Pasquale. "Anthropic Launches Managed Agents and Claude Cowork GA: The Triple Announcement of April 9, 2026." 2026. https://pasqualepillitteri.it/en/news/755/anthropic-managed-agents-cowork-ga-april-9-2026
36. InfoQ. "Anthropic Launches Claude Platform on AWS." May 2026. https://www.infoq.com/news/2026/05/anthropic-claude-aws/
37. Morphllm. "SWE-Bench Pro Leaderboard (2026): Why 46% Beats 81%." 2026. https://www.morphllm.com/swe-bench-pro
38. BenchLM. "TAU-bench Benchmark 2026: 38 model averages." Retrieved May 2026. https://benchlm.ai/benchmarks/tauBench
39. BenchLM. "Claude Sonnet 4.6 vs GPT-5.5: AI Benchmark Comparison 2026." 2026. https://benchlm.ai/compare/claude-sonnet-4-6-vs-gpt-5-5
40. Kilo.ai blog. "Benchmarking the Benchmarks: New GPT and Claude Releases Continue to One-Up Themselves." May 2026. https://blog.kilo.ai/p/benchmarking-the-benchmarks-new-gpt
41. Kim, Sean. "Microsoft 365 Copilot Now Runs Claude Sonnet: How 400 Million Office Users Got Access to Anthropic's AI." March 2026. https://blog.imseankim.com/microsoft-365-copilot-claude-sonnet-anthropic-enterprise-integration-2026/
42. Microsoft Copilot Blog. "Available today: Claude Sonnet 4.5 in Microsoft Copilot Studio." Microsoft, 2025. https://www.microsoft.com/en-us/microsoft-copilot/blog/copilot-studio/available-today-claude-sonnet-4-5-in-microsoft-copilot-studio/
43. Snowflake / Anthropic. "Snowflake and Anthropic Announce $200 Million Partnership to Bring Agentic AI to Global Enterprises." Snowflake press release, December 3, 2025. https://www.snowflake.com/en/news/press-releases/snowflake-and-anthropic-announce-200-million-partnership-to-bring-agentic-ai-to-global-enterprises/
44. Winbuzzer. "Microsoft Shifts Engineers from Claude Code to GitHub Copilot CLI." May 15, 2026. https://winbuzzer.com/2026/05/15/microsoft-starts-canceling-claude-code-licenses-xcxwbn/
45. Demandsage. "Claude Statistics 2026: 18.9M Active Users & $14B Revenue." 2026. https://www.demandsage.com/claude-ai-statistics/
46. Presenc. "Claude Usage Statistics 2026." Retrieved May 2026. https://presenc.ai/research/claude-usage-statistics
47. GetPanto. "Anthropic AI Statistics 2026: Users, Revenue & Market Share." 2026. https://www.getpanto.ai/blog/anthropic-ai-statistics
48. Salesforce Developers. "New in Salesforce Developer Edition: Agentforce Vibes IDE, Claude 4.5, MCP." April 2026. https://developer.salesforce.com/blogs/2026/04/new-developer-edition-agentforce-vibes-claude-mcp
49. NxCode. "Claude Sonnet 4.8 (Not 4.7): Release Date, Features & What to Expect (2026)." 2026. https://www.nxcode.io/resources/news/claude-sonnet-4-8-release-date-features-what-to-expect-2026
50. Geeky-Gadgets. "Claude Sonnet 4.8 Leaks Reveal About Anthropic's Next AI Release." 2026. https://www.geeky-gadgets.com/claude-sonic-4-8-leaks/
51. Microsoft. "Microsoft 365 E7 and Agent 365 are now generally available." Microsoft 365 Blog, May 1, 2026. https://techcommunity.microsoft.com/blog/microsoft_365blog/microsoft-365-e7-and-agent-365-are-now-generally-available/4516295
52. Althoff, Judson. "Introducing the First Frontier Suite built on Intelligence + Trust." The Official Microsoft Blog, March 9, 2026. https://blogs.microsoft.com/blog/2026/03/09/introducing-the-first-frontier-suite-built-on-intelligence-trust/
53. Anthropic. "The advisor strategy: Give agents an intelligence boost." Claude blog, April 9, 2026. https://claude.com/blog/the-advisor-strategy
54. Anthropic. "Claude Sonnet 4.6." Claude Platform Docs (legacy model page). Retrieved November 2026. https://platform.claude.com/docs/en/models/sonnet-4-6/overview