# Claude 3.5 Haiku

> Source: https://aiwiki.ai/wiki/claude_3_5_haiku
> 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 3.5 Haiku." aiwiki.ai, 28 Oct 2026. https://aiwiki.ai/wiki/claude_3_5_haiku
> From AI Wiki (https://aiwiki.ai), the free encyclopedia of artificial intelligence. Reuse freely with attribution.

**Claude 3.5 Haiku** is a fast, low-cost [large language model](https://aiwiki.ai/wiki/large_language_model) developed by [Anthropic](https://aiwiki.ai/wiki/anthropic), announced on October 22, 2024 and released through the [Claude](https://aiwiki.ai/wiki/claude) API on November 4, 2024 as the smallest and most affordable member of the Claude 3.5 generation. It is built for low-latency, high-volume tasks such as user-facing assistants, sub-agent routines, and large-scale data processing, and despite being an entry-tier model it matched or exceeded Anthropic's earlier flagship, [Claude 3 Opus](https://aiwiki.ai/wiki/claude_3_opus), on several benchmarks, including a reported 40.6% on [SWE-bench Verified](https://aiwiki.ai/wiki/swe-bench_verified).[1][2][3] Anthropic launched it with a 200,000-token [context window](https://aiwiki.ai/wiki/context_window) at a price of $1.00 per million input tokens and $5.00 per million output tokens, a fourfold increase over the original [Claude 3 Haiku](https://aiwiki.ai/wiki/claude_3_haiku) that drew sharp criticism, then cut the price to $0.80 / $4.00 per million tokens about a month later.[2][7][8][13] Anthropic retired the model from its API on February 19, 2026.[6]

<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>

## What is Claude 3.5 Haiku?

Claude 3.5 Haiku is the entry-tier model of Anthropic's Claude 3.5 family, positioned as a successor to the original Claude 3 Haiku and built for speed and cost efficiency rather than maximum reasoning power. Anthropic announced it on October 22, 2024, in the same blog post that introduced an upgraded version of [Claude 3.5 Sonnet](https://aiwiki.ai/wiki/claude_3_5_sonnet) and a public beta of the "[computer use](https://aiwiki.ai/wiki/computer_use)" capability. At announcement, the company said the model would arrive "later this month" across its first-party API, [Amazon Bedrock](https://aiwiki.ai/wiki/amazon_bedrock), and [Google Cloud](https://aiwiki.ai/wiki/google_cloud)'s Vertex AI, and that it would launch as a text-only model with image input to follow.[1][3] The model became available through the Claude API on November 4, 2024, and reached Amazon Bedrock the same day.[2][4][5]

The Claude naming scheme uses three size tiers: Haiku for the smallest and fastest models, Sonnet for the mid-range, and Opus for the largest. Claude 3.5 Haiku occupied the entry tier of the 3.5 family. Its API model identifier is `claude-3-5-haiku-20241022`, and Anthropic gave it a training-data cutoff of July 2024.[2][17]

## What can Claude 3.5 Haiku do?

Anthropic described Claude 3.5 Haiku as improving "across every skill set" relative to the previous Haiku while keeping similar speed, and highlighted gains in coding, instruction following, and tool use. The company suggested it for user-facing products, specialized sub-agent tasks within larger agentic systems, and generating personalized output from large volumes of data. In Anthropic's words, the model "is well suited for user-facing products, specialized sub-agent tasks, and generating personalized experiences from huge volumes of data, like purchase history, pricing, or inventory records."[1][3]

The model had a context window of up to 200,000 tokens and could generate up to 8,192 tokens in a single response.[2][7] At launch it processed text only. Unlike the original Claude 3 Haiku, which accepted images, the 3.5 version shipped without vision, and Anthropic advised users who needed image analysis or the lowest possible cost to continue using Claude 3 Haiku.[2][8] Vision support was added later: Anthropic's API release notes recorded image input arriving for Claude 3.5 Haiku on February 24, 2025, roughly four months after the text-only debut.[9]

## How does Claude 3.5 Haiku perform on benchmarks?

Anthropic's central claim was that, despite being the smallest model in the 3.5 generation, Claude 3.5 Haiku surpassed Claude 3 Opus on many intelligence benchmarks at a fraction of the cost and at the speed of the prior Haiku.[1][3] The most cited single figure was on coding: the company reported a score of 40.6% on SWE-bench Verified, a benchmark of real-world software-engineering tasks, and said this result outperformed "many agents using publicly available state-of-the-art models, including the original Claude 3.5 Sonnet and GPT-4o."[1][3][8]

Additional benchmark results reported for the model are summarized below. Scores are drawn from Anthropic's materials and aggregators that cite them; exact evaluation settings (shot count, [chain-of-thought](https://aiwiki.ai/wiki/chain_of_thought)) vary by benchmark.

| Benchmark | Domain | Claude 3.5 Haiku |
| --- | --- | --- |
| [SWE-bench Verified](https://aiwiki.ai/wiki/swe-bench_verified) | Software engineering | 40.6% |
| [HumanEval](https://aiwiki.ai/wiki/humaneval) | Code generation | 88.1% |
| [MMLU-Pro](https://aiwiki.ai/wiki/mmlu-pro) | General knowledge and reasoning | 65.0% |
| MATH | Mathematical problem solving | 69.2% |
| MGSM | Multilingual grade-school math | 85.6% |
| DROP | Reading comprehension | 83.1 (F1) |

Sources: Anthropic announcement for SWE-bench Verified; Anthropic's [model card](https://aiwiki.ai/wiki/model_card) addendum (Table 8) for the remaining figures.[1][17]

## How much does Claude 3.5 Haiku cost?

Pricing was the most discussed aspect of the release and went through three distinct stages. When Anthropic first previewed Claude 3.5 Haiku, it indicated that the model would cost the same as Claude 3 Haiku, which was priced at $0.25 per million input tokens and $1.25 per million output tokens.[8][11] At the actual API launch on November 4, 2024, the company instead set the price at $1.00 per million input tokens and $5.00 per million output tokens, a fourfold increase over Claude 3 Haiku on both figures.[2][8][11]

Anthropic justified the higher price by pointing to capability rather than parity with the older model. In a statement posted to its social accounts, the company wrote: "During final testing, Haiku surpassed Claude 3 Opus, our previous flagship model, on many benchmarks, at a fraction of the cost. As a result, we've increased pricing for Claude 3.5 Haiku to reflect its increase in intelligence."[8][12] About a month later, on December 3, 2024, Anthropic reduced the price by 20% to $0.80 per million input tokens and $4.00 per million output tokens, saying it was "lowering the price of Claude 3.5 Haiku" to make the model "even more accessible for a wide range of use cases." That lower rate became the model's standard pricing.[13][18]

| Date | Input (per 1M tokens) | Output (per 1M tokens) | Note |
| --- | --- | --- | --- |
| Pre-launch indication | $0.25 | $1.25 | Suggested parity with [Claude 3 Haiku](https://aiwiki.ai/wiki/claude_3_haiku) |
| November 4, 2024 (launch) | $1.00 | $5.00 | 4x increase over [Claude 3 Haiku](https://aiwiki.ai/wiki/claude_3_haiku) |
| December 3, 2024 | $0.80 | $4.00 | 20% reduction, standard rate thereafter |

The model was distributed through Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI.[1][2]

## Why was the Claude 3.5 Haiku price controversial?

Coverage of the launch concentrated on the pricing change rather than on the model's quality. TechCrunch noted that it is rare for an AI vendor to raise the cost of a model within a series and said the move raised questions about Anthropic's pricing strategy; the Digital Watch Observatory tied those questions to intensifying competition in the AI industry.[8][14] Several outlets framed the model as four times more expensive than its predecessor (TechRadar's headline called it "4 times more expensive than its predecessor"), and the Digital Watch Observatory noted that the missing image analysis might lead some users to stick with the cheaper Claude 3 Haiku.[8][11][14]

Reaction to the model itself was mixed. The strong SWE-bench Verified result drew attention from developers, and the model found use as a cheaper alternative to mid-tier models for coding and agent tasks where Opus-level reasoning was not required. Anthropic presented the 20% price cut in December as a way to make the model "even more accessible for a wide range of use cases."[13][18]

## What replaced Claude 3.5 Haiku?

Claude 3.5 Haiku was later succeeded in the Haiku tier by faster, more capable models in subsequent Claude generations, including [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5), which posted substantially higher coding scores while continuing the same emphasis on speed and low cost.[16] Anthropic notified developers on December 19, 2025 that the `claude-3-5-haiku-20241022` snapshot was deprecated, named Claude Haiku 4.5 as the recommended replacement, and retired the model from the Claude API on February 19, 2026.[6] As of November 2026 the current Haiku-tier model is [Claude Haiku 5.5](https://aiwiki.ai/wiki/claude_haiku_5_5), released on October 7, 2026.[19]

## ELI5

Think of Anthropic's Claude models as coming in three sizes: small and quick (Haiku), medium (Sonnet), and large and powerful (Opus). Claude 3.5 Haiku is the small, fast, cheap one. Even though it is the little model, it turned out to be smart enough to beat the previous biggest model (Claude 3 Opus) on a lot of tests, which surprised people. Because it was smarter than expected, Anthropic charged four times more for it than the old small model, everyone complained, and a month later Anthropic lowered the price a bit.

## See also

- [Claude](https://aiwiki.ai/wiki/claude)
- [Anthropic](https://aiwiki.ai/wiki/anthropic)
- [Claude 3 Haiku](https://aiwiki.ai/wiki/claude_3_haiku)
- [Claude 3.5 Sonnet](https://aiwiki.ai/wiki/claude_3_5_sonnet)
- [Claude 3 Opus](https://aiwiki.ai/wiki/claude_3_opus)
- [Claude Haiku 4.5](https://aiwiki.ai/wiki/claude_haiku_4_5)
- [Large language model](https://aiwiki.ai/wiki/large_language_model)

## References

1. [Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku](https://www.anthropic.com/news/3-5-models-and-computer-use). Anthropic, October 22, 2024.
2. [Claude 3.5 Haiku](https://simonwillison.net/2024/Nov/4/haiku/). Simon Willison, November 4, 2024.
3. [Anthropic just dropped Claude Haiku 3.5 and gave the chatbot a huge upgrade](https://www.tomsguide.com/ai/anthropic-just-dropped-claude-haiku-3-5-and-gave-the-chatbot-a-huge-upgrade-heres-whats-new). Tom's Guide.
4. [Anthropic's Claude 3.5 Haiku model now available in Amazon Bedrock](https://aws.amazon.com/about-aws/whats-new/2024/11/anthropics-claude-3-5-haiku-model-amazon-bedrock/). Amazon Web Services, November 2024.
5. [Amazon Bedrock now supports Claude 3.5 Haiku and upgraded Claude 3.5 Sonnet](https://aws.amazon.com/blogs/aws/upgraded-claude-3-5-sonnet-from-anthropic-available-now-computer-use-public-beta-and-claude-3-5-haiku-coming-soon-in-amazon-bedrock/). AWS Blog.
6. [Model deprecations](https://platform.claude.com/docs/en/about-claude/model-deprecations). Claude API documentation.
7. [Claude 3.5 Haiku model card](https://www.prompthub.us/models/claude-3-5-haiku). PromptHub.
8. [Anthropic hikes the price of its Haiku model](https://techcrunch.com/2024/11/04/anthropic-hikes-the-price-of-its-haiku-model/). TechCrunch, November 4, 2024.
9. [Claude 3.5 Haiku has vision now](https://github.com/simonw/llm-anthropic/issues/17). llm-anthropic GitHub issue, citing Anthropic API release notes, February 2025.
10. [Claude 3.5 Haiku benchmarks](https://llm-stats.com/models/claude-3-5-haiku-20241022). LLM Stats.
11. [Anthropic's new Claude 3.5 Haiku AI model is 4 times more expensive than its predecessor](https://www.techradar.com/computing/artificial-intelligence/anthropics-new-claude-3-5-haiku-ai-model-is-4-times-more-expensive-than-its-predecessor). TechRadar.
12. [Anthropic statement on Claude 3.5 Haiku pricing](https://x.com/AnthropicAI/status/1853498270724542658). Anthropic, November 2024.
13. [Claude 3.5 Haiku price drops by 20%](https://simonwillison.net/2024/Dec/5/claude-35-haiku-price-drops-by-20/). Simon Willison, December 5, 2024.
14. [Anthropic launches Claude 3.5 Haiku with price rise and feature trade-offs](https://dig.watch/updates/anthropic-launches-claude-3-5-haiku-with-price-rise-and-feature-trade-offs-amid-ai-competition). Digital Watch Observatory.
15. [Anthropic releases updated, smarter Claude Haiku 3.5 and Sonnet 3.5 model](https://the-decoder.com/anthropic-releases-updated-smarter-claude-haiku-3-5-and-sonnet-3-5-model/). The Decoder.
16. [Introducing Claude Haiku 4.5](https://www.anthropic.com/news/claude-haiku-4-5). Anthropic.
17. [Model Card Addendum: Claude 3.5 Haiku and Upgraded Claude 3.5 Sonnet](https://assets.anthropic.com/m/1cd9d098ac3e6467/original/Claude-3-Model-Card-October-Addendum.pdf). Anthropic, October 2024.
18. [Claude 3.5 Haiku on AWS Trainium2 and model distillation in Amazon Bedrock](https://www.anthropic.com/news/trainium2-and-distillation). Anthropic, December 3, 2024.
19. [Introducing Claude Haiku 5.5](https://www.anthropic.com/claude-haiku-5-5). Anthropic, October 7, 2026.

