116 reviewed model records

AI model comparison finder

Filter by your constraints, then compare context, pricing, and lifecycle side by side, each linked to a first-party source.

No sponsored rankings · No login or API key · Reviewed Oct 25, 2026, today

116 comparison records · Browse all model articles

This tool uses manually reviewed pricing assumptions with the dates shown below. For the separately refreshed price snapshot and named intelligence benchmarks, see API prices on the front page. Check the source and rate configuration before using an estimate.

Models you add below carry over to the cost, lifecycle, and memory tools.

Side by side

Compare selected models

Suggested start: the newest flagship from OpenAI, Anthropic, Google. The first model you add replaces it; nothing is saved until you choose.

Only show differences

3 of 4 selected

Side-by-side comparison of GPT-6.1 Sol, Claude Sonnet 5.5, Gemini 4 Argon
Attribute	OpenAI GPT-6.1 Sol	Anthropic Claude Sonnet 5.5	Google Gemini 4 Argon
Lifecycle	GPT-6.1 Sol Active No retirement announced	Claude Sonnet 5.5 Active Available until at least Sep 28, 2027	Gemini 4 Argon Active No retirement announced
API / checkpoint ID	GPT-6.1 Sol `gpt-6.1-sol`	Claude Sonnet 5.5 `claude-sonnet-5-5`	Gemini 4 Argon `gemini-4-argon`
Release date	GPT-6.1 Sol Sep 29, 2026	Claude Sonnet 5.5 Sep 28, 2026	Gemini 4 Argon Oct 13, 2026
Knowledge cutoff	GPT-6.1 Sol April 30, 2026	Claude Sonnet 5.5 June 2026	Gemini 4 Argon Not published
Distribution	GPT-6.1 Sol Hosted API	Claude Sonnet 5.5 Hosted API	Gemini 4 Argon Hosted API
Context window	GPT-6.1 Sol 1.05M tokensLargest	Claude Sonnet 5.5 1M tokens	Gemini 4 Argon 1.04M tokens
Maximum output	GPT-6.1 Sol 128K tokens	Claude Sonnet 5.5 128K tokens	Gemini 4 Argon 65.5K tokens
Modalities	GPT-6.1 Sol text, image to text	Claude Sonnet 5.5 text, image, pdf to text	Gemini 4 Argon text, image, video, audio, pdf to text
Capabilities	GPT-6.1 Sol Reasoning / thinkingTool use / function callingStructured outputsPrompt / context cachingBatch processing Not supported: Self-hosting Unknown: Code execution, Search grounding, Fine-tuning, Multilingual	Claude Sonnet 5.5 Reasoning / thinkingTool use / function callingStructured outputsPrompt / context cachingBatch processing Not supported: Self-hosting Unknown: Code execution, Search grounding, Fine-tuning, Multilingual	Gemini 4 Argon Reasoning / thinkingTool use / function callingStructured outputsCode executionSearch groundingPrompt / context cachingBatch processing Not supported: Self-hosting Unknown: Fine-tuning, Multilingual
Intelligence Index (third party)	GPT-6.1 Sol 51.8 Artificial Analysis Intelligence Index (via OpenRouter) Score checked Oct 25, 2026 Benchmark source (opens in a new tab) About this score Higher is better, not an IQ or percentage. Catalogue metadata does not disclose the index version, reasoning configuration, or evaluation date; checked time is retrieval time. Treat comparisons as indicative.	Claude Sonnet 5.5 56 Artificial Analysis Intelligence Index (via OpenRouter) Score checked Oct 25, 2026 Benchmark source (opens in a new tab) About this score Higher is better, not an IQ or percentage. Catalogue metadata does not disclose the index version, reasoning configuration, or evaluation date; checked time is retrieval time. Treat comparisons as indicative.	Gemini 4 Argon 52.6 Artificial Analysis Intelligence Index (via OpenRouter) Score checked Oct 25, 2026 Benchmark source (opens in a new tab) About this score Higher is better, not an IQ or percentage. Catalogue metadata does not disclose the index version, reasoning configuration, or evaluation date; checked time is retrieval time. Treat comparisons as indicative.
Standard text price	GPT-6.1 Sol $2 input / $10 output USD per 1M tokens Price checked 2026-10-24 Higher rates above 272K input tokens ($4 input / $15 output). Cached input is billed at 5 percent of the input rate, half of GPT-6 Sol's cached price. Flex is priced like Batch, Fast mode costs 2x Standard, and regional processing adds 10 percent where available.	Claude Sonnet 5.5 $2 input / $10 output USD per 1M tokens Price checked 2026-10-24 Base Claude API price. Cache reads were $0.20 per million tokens at launch on September 28, 2026 and were cut to $0.10 (0.05x the base input price) on October 7, 2026; all other rates are unchanged since launch. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.	Gemini 4 Argon $2 input / $10 output USD per 1M tokens Price checked 2026-10-24 $2/$10 through Nov 30, 2026, then $4/$20 Introductory paid-tier Standard rates through November 30, 2026; there is no free tier and only implicit caching. The output price includes thinking tokens.
Open-weight details	GPT-6.1 Sol Hosted API only	Claude Sonnet 5.5 Hosted API only	Gemini 4 Argon Hosted API only
Caveats	GPT-6.1 Sol Does not accept the none or minimal reasoning efforts, and tool calling requires the Responses API (Chat Completions works without tools).	Claude Sonnet 5.5 Forced tool use (tool_choice any or tool) returns a 400 error, and its thinking blocks work only in the account that produced them or a linked account. The Message Batches API allows up to 300K output tokens with the output-300k-2026-03-24 beta header.	Gemini 4 Argon Google publishes no knowledge cutoff for Gemini 4 Argon. Its 65,536-token output limit is the default for synchronous requests; Google lists up to 1,048,576 output tokens with long decode.
Official source	GPT-6.1 Sol GPT-6.1 Sol model documentation	Claude Sonnet 5.5 Claude Sonnet 5.5 model page	Gemini 4 Argon Gemini 4 Argon model documentation

Matching models

82 of 116 reviewed records

Sort

Meta

Muse Spark Open

muse-spark-open

Active

Context: 1.04M
Max output: Not published
Distribution: Open weights
Input / output price: Varies by host
Released: Oct 15, 2026
Knowledge cutoff: Not published

Not served on Meta's Model API: self-host the BF16 or NVFP4 weights or call a hosted inference provider. Meta lists at least 700GB of GPU memory for NVFP4 and 2TB for BF16. The license is Meta's custom Muse Model License, not an OSI license.

No retirement announced

text inputimage inputvideo inputaudio inputReasoning / thinkingTool use / function callingFine-tuning

Official source

Google

Gemini 4 Argon

gemini-4-argon

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $2 / $10
Released: Oct 13, 2026
Knowledge cutoff: Not published

$2/$10 through Nov 30, 2026, then $4/$20

Introductory paid-tier Standard rates through November 30, 2026; there is no free tier and only implicit caching. The output price includes thinking tokens.

Google publishes no knowledge cutoff for Gemini 4 Argon. Its 65,536-token output limit is the default for synchronous requests; Google lists up to 1,048,576 output tokens with long decode.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Haiku 5.5

claude-haiku-5-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $0.10 / $0.50
Released: Oct 7, 2026
Knowledge cutoff: June 2026

Higher rates above 100K input tokens ($0.50 input / $2.50 output).

Base Claude API price for prompts up to 100,000 tokens. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.

Unlike other current Claude models, Haiku 5.5 is priced by prompt length: prompts over 100,000 tokens cost $0.50 input and $2.50 output per million tokens. Manual extended thinking (budget_tokens) returns a 400 error.

Available until at least Oct 7, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-6.1 Sol

gpt-6.1-sol

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $2 / $10
Released: Sep 29, 2026
Knowledge cutoff: April 30, 2026

Higher rates above 272K input tokens ($4 input / $15 output).

Cached input is billed at 5 percent of the input rate, half of GPT-6 Sol's cached price. Flex is priced like Batch, Fast mode costs 2x Standard, and regional processing adds 10 percent where available.

Does not accept the none or minimal reasoning efforts, and tool calling requires the Responses API (Chat Completions works without tools).

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Sonnet 5.5

claude-sonnet-5-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $2 / $10
Released: Sep 28, 2026
Knowledge cutoff: June 2026

Base Claude API price. Cache reads were $0.20 per million tokens at launch on September 28, 2026 and were cut to $0.10 (0.05x the base input price) on October 7, 2026; all other rates are unchanged since launch. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.

Forced tool use (tool_choice any or tool) returns a 400 error, and its thinking blocks work only in the account that produced them or a linked account. The Message Batches API allows up to 300K output tokens with the output-300k-2026-03-24 beta header.

Available until at least Sep 28, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 5.5

claude-opus-5-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $4 / $20
Released: Sep 22, 2026
Knowledge cutoff: June 2026

Base Claude API price. Cache reads cost 0.05x the base input price on this model. Fast mode (out of research preview since October 15, 2026) costs $8 input and $40 output per million tokens. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.

Adaptive thinking is always on and cannot be turned off, and forced tool use (tool_choice any or tool) returns a 400 error. The Message Batches API allows up to 300K output tokens with the output-300k-2026-03-24 beta header.

Available until at least Sep 22, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-6 Luna

gpt-6-luna

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $0.10 / $0.50
Released: Sep 22, 2026
Knowledge cutoff: May 18, 2026

Higher rates above 272K input tokens ($0.20 input / $0.75 output).

Flex is priced like Batch, Fast mode costs 2x Standard, and regional processing adds 10 percent where available.

On September 25, 2026 OpenAI fixed an image-encoding bug that had degraded image understanding, so rerun image evaluations made before then. Chat Completions supports function calling only with reasoning effort none; use the Responses API for tools with reasoning.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-6 Sol

gpt-6-sol

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $2 / $10
Released: Sep 22, 2026
Knowledge cutoff: April 20, 2026

Higher rates above 272K input tokens ($4 input / $15 output).

Flex is priced like Batch, Fast mode costs 2x Standard, and regional processing adds 10 percent where available.

OpenAI's docs now point to GPT-6.1 Sol as the newer Sol model, though gpt-6-sol has no announced deprecation. On September 25, 2026 OpenAI fixed an image-encoding bug that had degraded image understanding, so rerun image evaluations made before then.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.7

grok-4.7

Active

Context: 500K
Max output: Not published
Distribution: Hosted API
Input / output price: $2 / $6
Released: Sep 21, 2026
Knowledge cutoff: May 2026

Higher rates at or above 200K input tokens ($4 input / $12 output).

xAI lists a 20% Batch API discount for this model (on all token types); batch rates are not yet in this registry. Requests to the US regional endpoint (https://us.api.x.ai/v1) bill at 1.1x these rates, Priority Processing bills at 2x, and server-side tool calls are billed separately.

xAI's model page lists no maximum output limit and xAI publishes no retirement date. Batch API requests are supported in us-east-1 and us-west-2 at a 20% discount to standard rates. Grok 4.7 Fast (the same model at 2x standard rates, or 1.5x for long-context requests) is sold only in Cursor and Grok Build, not on the public xAI API.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

DeepSeek

DeepSeek-V4.1-Flash

deepseek-flash

Active

Context: 1M
Max output: 384K
Distribution: API and downloadable weights
Input / output price: $0.30 / $1.20
Released: Sep 10, 2026
Knowledge cutoff: Not published

Peak rates are the list rates. Off-peak input/cache/output: $0.15/$0.003/$0.60 (half the peak rates). Peak hours are 01:00-04:00 and 06:00-10:00 UTC, Monday through Friday, excluding Chinese public holidays; weekends and Chinese public holidays are off-peak all day. The cost planner blends peak and off-peak rates by the share of traffic sent in peak hours (35 of 168 weekly hours when spread evenly).

API model ID is deepseek-flash; the retired names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, served by this model and billed at these rates. Since October 17, 2026 the API serves the DeepSeek-V4.1-Flash-1017 build, with the same model name and prices. Generic dense-transformer memory formulas do not model this architecture accurately.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-6 Astra

gpt-6-astra

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $10 / $50
Released: Sep 3, 2026
Knowledge cutoff: April 30, 2026

Higher rates above 272K input tokens ($20 input / $75 output).

Flex is priced like Batch and Fast mode costs 2x Standard. Since September 29, 2026 an Ultrafast tier (service_tier ultrafast, Responses API) costs $60 input, $6 cached input, $75 cache writes and $300 output per 1M tokens, or $120, $12, $150 and $450 above 272K input tokens.

Does not accept the none reasoning effort, custom temperature or top_p values, or log probabilities, and tool calling requires the Responses API (Chat Completions works without tools).

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 3.8 Flash

gemini-3.8-flash

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $0.75 / $3.75
Released: Sep 2, 2026
Knowledge cutoff: March 2026

$0.75/$3.75 through Dec 31, 2026, then $1.50/$7.50

Introductory paid-tier Standard rates announced at launch and valid through December 31, 2026; the output price includes thinking tokens. Cache storage ($0.50 per 1M tokens per hour) and grounding are billed separately.

Google qualifies the March 2026 knowledge cutoff: some domains may be limited to January 2025. Google warns the model may use more tokens on complex tasks, especially at higher effort levels.

No retirement announced

text inputimage inputaudio inputvideo inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Meta

Muse Spark 1.3

muse-spark-1.3

Active

Context: 1.04M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.25 / $4.25
Released: Sep 2, 2026
Knowledge cutoff: Not published

Standard tier rates as observed on October 8, 2026 (Meta's pricing page does not date them); prompts and completions are not used to train Meta models. No long-context premium; built-in web search grounding adds $2.50 per 1,000 queries. The muse-spark-1.3-contributor variant bills $0.10/$0.002/$0.20 in exchange for training use.

Meta lists audio input but says audio understanding in Muse Spark 1.3 is not fully supported and may be degraded (it points audio work to Muse Spark 1.2 or Muse Voice Transcribe). Meta does not publish a maximum output limit, knowledge cutoff, or parameter count, and the max reasoning level is Standard-tier only.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Fable 5.1

claude-fable-5-1

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $10 / $50
Released: Sep 1, 2026
Knowledge cutoff: June 2026

Base Claude API price. Cache reads cost 0.025x the base input price on this model. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.

Anthropic requires 30-day data retention for this model; it is not available under zero data retention unless Anthropic expressly authorizes it. Forced tool use (tool_choice any or tool) returns a 400 error.

Available until at least Sep 1, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Mythos 5.1

claude-mythos-5-1

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $10 / $50
Released: Sep 1, 2026
Knowledge cutoff: June 2026

Anthropic lists the same specifications and pricing as Claude Fable 5.1, including cache reads at 0.025x the base input price.

Not generally available: Anthropic offers it only to organizations verified through its verification programs, such as the Cyber Verification Program (whose three access tiers, announced October 6, 2026, each include it) and the Life Sciences Verification Program; at launch it was limited to a set of US organizations. It is the same model as Claude Fable 5.1 with different safeguards, and it carries 30-day data retention unless Anthropic expressly authorizes zero data retention.

Available until at least Sep 1, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.6

grok-4.6

Active

Context: 500K
Max output: Not published
Distribution: Hosted API
Input / output price: $2 / $6
Released: Aug 12, 2026
Knowledge cutoff: February 1, 2026

Higher rates at or above 200K input tokens ($4 input / $12 output).

xAI's model page lists no maximum output limit and xAI publishes no retirement date. Batch API requests are supported in us-east-1 and us-west-2 at a 20% discount to standard rates. The models page recommends Grok 4.7, but grok-4.6 keeps its model page and its pricing row.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 5

claude-opus-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $5 / $25
Released: Jul 24, 2026
Knowledge cutoff: May 2026

Base Claude API price, the same as Claude Opus 4.8. Fast mode (research preview, Claude API only) costs $10 input and $50 output per million tokens. US-only inference (inference_geo) costs 1.1x, and partner-platform prices can differ.

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; its model page labels it legacy and recommends migrating to Claude Opus 5.5. Disabling thinking is allowed only at effort high or below.

Available until at least Jul 24, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 3.5 Flash-Lite

gemini-3.5-flash-lite

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $0.30 / $2.50
Released: Jul 21, 2026
Knowledge cutoff: Not published

Paid-tier text/image/video input price; feature charges are separate. Context caching is paid-tier only (plus $1.00 per 1M tokens per hour of cache storage).

The current Gemini API deprecation table has no shutdown date for this model. Model and service limits remain subject to the linked documentation.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 3.6 Flash

gemini-3.6-flash

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $0.75 / $3.75
Released: Jul 21, 2026
Knowledge cutoff: Not published

$0.75/$3.75 through Dec 31, 2026, then $1.50/$7.50

Google's Gemini 3.7 Flash guide says the 3.7 introductory rate also applies to 3.6 Flash. The first dated listing is the 3.7 Flash model card published August 13, 2026; Google states no separate start date. Launch list price was $1.50 input and $7.50 output. Cache storage and grounding are billed separately.

The current Gemini API deprecation table has no shutdown date for this model. Model and service limits remain subject to the linked documentation.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.5

grok-4.5

Active

Context: 500K
Max output: Not published
Distribution: Hosted API
Input / output price: $2 / $6
Released: Jul 16, 2026
Knowledge cutoff: February 1, 2026

Higher rates at or above 200K input tokens ($4 input / $12 output).

The official model page lists no maximum output limit or committed retirement date. Alias mappings can change; use the explicit model ID.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.6 Luna

gpt-5.6-luna

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $0.20 / $1.20
Released: Jul 9, 2026
Knowledge cutoff: February 16, 2026

Higher rates above 272K input tokens ($0.40 input / $1.80 output).

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.6 Sol

gpt-5.6-sol

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $4 / $20
Released: Jul 9, 2026
Knowledge cutoff: February 16, 2026

Higher rates above 272K input tokens ($8 input / $30 output).

Promotional rates are available at least through November 21, 2026. Later rates have not been committed; projections hold these rates constant.

The gpt-5.6 alias currently resolves to Sol; aliases are mutable, so pin gpt-5.6-sol when reproducibility matters.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.6 Terra

gpt-5.6-terra

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $2 / $12
Released: Jul 9, 2026
Knowledge cutoff: February 16, 2026

Higher rates above 272K input tokens ($4 input / $18 output).

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Sonnet 5

claude-sonnet-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $2 / $10
Released: Jun 30, 2026
Knowledge cutoff: January 2026

These are standard prices. The previously announced September 1 increase was cancelled.

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; its model page labels it legacy and recommends Claude Sonnet 5.5.

Available until at least Jun 30, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Fable 5

claude-fable-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $10 / $50
Released: Jun 9, 2026
Knowledge cutoff: January 2026

Base Claude API price; regional and partner-platform charges can differ.

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; its model page labels it legacy and recommends Claude Fable 5.1. US export controls forced a global suspension of Fable 5 and Mythos 5 from June 12 to June 30, 2026; access was restored on July 1, 2026.

Available until at least Jun 9, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Mythos 5

claude-mythos-5

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $10 / $50
Released: Jun 9, 2026
Knowledge cutoff: January 2026

Anthropic lists the same specs and pricing as Claude Fable 5.

Not generally available: Anthropic offers it only to organizations verified through its verification programs, such as the Cyber Verification Program, which Anthropic merged with Project Glasswing into one expanded program on October 6, 2026; Claude Mythos 5.1 is now the current Mythos model. Also covered by the June 12 to June 30, 2026 export-control suspension; access was restored July 1, 2026.

Available until at least Jun 9, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok Build 0.1

grok-build-0.1

Active

Context: 256K
Max output: Not published
Distribution: Hosted API
Input / output price: $1 / $2
Released: May 29, 2026
Knowledge cutoff: Not published

Higher rates at or above 200K input tokens ($2 input / $4 output).

xAI announced grok-build-0.1 on its API in public beta on May 29, 2026 (its May release notes call it early access). The official model page lists no maximum output limit or committed retirement date, and alias mappings can change, so use the explicit model ID.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 4.8

claude-opus-4-8

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $5 / $25
Released: May 28, 2026
Knowledge cutoff: January 2026

Base Claude API price; regional and partner-platform charges can differ.

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; its model page labels it legacy and recommends Claude Opus 5.5.

Available until at least May 28, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Mistral AI

Mistral Medium 3.5

mistral-medium-3-5

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $1.50 / $7.50
Released: Apr 28, 2026
Knowledge cutoff: Not published

First-party Mistral API price. Cached-input and batch rates are published directly on Mistral's pricing page (Standard and Batch views).

Mistral publishes dateless IDs only (mistral-medium-3-5 plus the mistral-medium-3 and mistral-medium-latest aliases); the release date and license wording come from the model card alone.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

DeepSeek

DeepSeek-V4-Pro

deepseek-v4-pro

Active

Context: 1M
Max output: 384K
Distribution: API and downloadable weights
Input / output price: $1.32 / $3.96
Released: Apr 24, 2026
Knowledge cutoff: Not published

Peak rates are the list rates. Off-peak input/cache/output: $0.66/$0.022/$1.98 (half the peak rates). Peak hours are 01:00-04:00 and 06:00-10:00 UTC, Monday through Friday, excluding Chinese public holidays; weekends and Chinese public holidays are off-peak all day. The cost planner blends peak and off-peak rates by the share of traffic sent in peak hours (35 of 168 weekly hours when spread evenly).

Served by the DeepSeek-V4-Pro-0813 build since August 13, 2026; image input is not supported. DeepSeek's September 10 news post still says this name routes to V4.1-Flash from September 14, but its change log says V4-Pro API service continues after that date with billing unchanged until further notice, and the pricing page still lists it.

No retirement announced

text inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 4.7

claude-opus-4-7

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $5 / $25
Released: Apr 16, 2026
Knowledge cutoff: January 2026

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; the models overview groups it under legacy models. Fast mode was deprecated June 25, 2026, with removal on July 24, 2026.

Available until at least Apr 16, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemma 4 31B

gemma-4-31b-it

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: Varies by host
Released: Apr 2, 2026
Knowledge cutoff: Not published

Open-weight hosting has no universal price or output cap. The first-party Gemini API deployment can have separate service limits.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingFine-tuning

Official source

Mistral AI

Mistral Small 4

mistral-small-2603

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.15 / $0.60
Released: Mar 16, 2026
Knowledge cutoff: Not published

First-party Mistral API price; self-hosting cost is deployment-specific. Cached-input and batch rates are published directly on Mistral's pricing page (Standard and Batch views).

mistral-small-latest is a rolling alias. The published model card does not specify a separate maximum-output limit.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Sonnet 4.6

claude-sonnet-4-6

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $3 / $15
Released: Feb 17, 2026
Knowledge cutoff: August 2025

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; the models overview groups it under legacy models.

Available until at least Feb 17, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 4.6

claude-opus-4-6

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $5 / $25
Released: Feb 5, 2026
Knowledge cutoff: May 2025

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; the models overview groups it under legacy models.

Available until at least Feb 5, 2027

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Mistral AI

Ministral 3 14B

ministral-14b-2512

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.20 / $0.20
Released: Dec 2, 2025
Knowledge cutoff: Not published

Mistral's Batch rates were not reviewed.

Instruct, reasoning and base weights are published under Apache 2.0. Mistral publishes no maximum output limit.

No retirement announced

text inputimage inputTool use / function callingStructured outputsPrompt / context caching

Official source

Mistral AI

Ministral 3 3B

ministral-3b-2512

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.10 / $0.10
Released: Dec 2, 2025
Knowledge cutoff: Not published

Mistral's Batch rates were not reviewed.

Instruct, reasoning and base weights are published under Apache 2.0. Mistral publishes no maximum output limit.

No retirement announced

text inputimage inputTool use / function callingStructured outputsPrompt / context caching

Official source

Mistral AI

Ministral 3 8B

ministral-8b-2512

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.15 / $0.15
Released: Dec 2, 2025
Knowledge cutoff: Not published

Mistral's Batch rates were not reviewed.

Instruct, reasoning and base weights are published under Apache 2.0. Mistral publishes no maximum output limit.

No retirement announced

text inputimage inputTool use / function callingStructured outputsPrompt / context caching

Official source

Mistral AI

Mistral Large 3

mistral-large-2512

Active

Context: 256K
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.50 / $1.50
Released: Dec 2, 2025
Knowledge cutoff: Not published

First-party Mistral API price. Cached-input and batch rates are published directly on Mistral's pricing page (Standard and Batch views).

mistral-large-latest is a rolling alias that still points to Mistral Large 3 (mistral-large-2512), not to the Mistral Large 4 public preview. The published model card does not specify a separate maximum-output limit.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Opus 4.5

claude-opus-4-5-20251101

Active

Context: 200K
Max output: 64K
Distribution: Hosted API
Input / output price: $5 / $25
Released: Nov 24, 2025
Knowledge cutoff: May 2025

Anthropic's lifecycle table lists this model as active with a not-sooner-than commitment; the models overview groups it under legacy models.

Available until at least Nov 24, 2026

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Anthropic

Claude Haiku 4.5

claude-haiku-4-5-20251001

Active

Context: 200K
Max output: 64K
Distribution: Hosted API
Input / output price: $1 / $5
Released: Oct 15, 2025
Knowledge cutoff: February 2025

Base Claude API price; regional and partner-platform charges can differ.

The dated API ID is pinned; claude-haiku-4-5 is a convenience alias. Anthropic's not-sooner-than date of October 15, 2026 has passed, and on October 25, 2026 its lifecycle table still listed the model as active with no deprecation notice. Its model page labels it legacy and recommends Claude Haiku 5.5.

The availability commitment ended Oct 15, 2026; no retirement date announced

text inputimage inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

gpt-oss-120b

gpt-oss-120b

Active

Context: 131K
Max output: 131K
Distribution: Open weights
Input / output price: Varies by host
Released: Aug 5, 2025
Knowledge cutoff: Not published

Open weights have no universal token price. Hosting cost, quantization, context limits, and throughput depend on the deployment.

No retirement announced

text inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

gpt-oss-20b

gpt-oss-20b

Active

Context: 131K
Max output: 131K
Distribution: Open weights
Input / output price: Varies by host
Released: Aug 5, 2025
Knowledge cutoff: Not published

Open weights have no universal token price. Hosting cost, quantization, context limits, and throughput depend on the deployment.

No retirement announced

text inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Mistral AI

Codestral 25.08

codestral-2508

Active

Context: 128K
Max output: Not published
Distribution: Hosted API
Input / output price: $0.30 / $0.90
Released: Jul 30, 2025
Knowledge cutoff: Not published

Mistral's Batch rates were not reviewed.

Mistral's code completion model (FIM). Mistral publishes no maximum output limit.

No retirement announced

text inputTool use / function callingStructured outputsPrompt / context caching

Official source

Google

Gemini 2.5 Flash-Lite

gemini-2.5-flash-lite

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $0.10 / $0.40
Released: Jul 22, 2025
Knowledge cutoff: Not published

Not deprecated: Google's deprecation table lists no shutdown date and says the 2.5 models will be served until further notice. Since September 18, 2026 Google limits access to users who have actively used them before, and recommends 3.5 Flash-Lite or 3.8 Flash for new projects.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 2.5 Flash

gemini-2.5-flash

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $0.30 / $2.50
Released: Jun 17, 2025
Knowledge cutoff: Not published

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 2.5 Pro

gemini-2.5-pro

Active

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $1.25 / $10
Released: Jun 17, 2025
Knowledge cutoff: Not published

Higher rates above 200K input tokens ($2.50 input / $15 output).

Paid-tier price for prompts up to 200K tokens. Longer prompts use higher input and output rates.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Meta

Llama 4 Maverick

llama-4-maverick

Active

Context: 1M
Max output: Not published
Distribution: Open weights
Input / output price: Varies by host
Released: Apr 5, 2025
Knowledge cutoff: Not published

Llama is open-weight under Meta's custom community license, not OSI open source. Hosted providers can impose different limits.

No retirement announced

text inputimage inputFine-tuningSelf-hostingMultilingual

Official source

Meta

Llama 4 Scout

llama-4-scout

Active

Context: 10M
Max output: Not published
Distribution: Open weights
Input / output price: Varies by host
Released: Apr 5, 2025
Knowledge cutoff: Not published

Llama is open-weight under Meta's custom community license, not OSI open source. Hosted providers may expose a smaller context window.

No retirement announced

text inputimage inputFine-tuningSelf-hostingMultilingual

Official source

Amazon

Amazon Nova Lite

amazon.nova-lite-v1:0

Active

Context: 300K
Max output: 10K
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Amazon Nova version 1, served through Amazon Bedrock (the alias is the US cross-region inference profile). AWS says Amazon Nova 2 is now available; Nova 2 was not reviewed. Bedrock prices were not reviewed.

No retirement announced

text inputimage inputvideo inputpdf inputBatch processingFine-tuning

Official source

Amazon

Amazon Nova Micro

amazon.nova-micro-v1:0

Active

Context: 128K
Max output: 10K
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

No retirement announced

text inputBatch processingFine-tuning

Official source

Amazon

Amazon Nova Premier

amazon.nova-premier-v1:0

Active

Context: 1M
Max output: 10K
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

No retirement announced

text inputimage inputvideo inputpdf inputBatch processing

Official source

Amazon

Amazon Nova Pro

amazon.nova-pro-v1:0

Active

Context: 300K
Max output: 10K
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

No retirement announced

text inputimage inputvideo inputpdf inputBatch processingFine-tuning

Official source

Cohere

Command A

command-a-03-2025

Active

Context: 256K
Max output: 8,000
Distribution: API and downloadable weights
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Built for tool use, agents, retrieval-augmented generation and multilingual use. Cohere states context and output limits in rounded thousands; prices were not reviewed.

No retirement announced

text inputTool use / function callingSelf-hostingMultilingual

Official source

Cohere

Command A Reasoning

command-a-reasoning-08-2025

Active

Context: 256K
Max output: 32K
Distribution: API and downloadable weights
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Cohere's first reasoning model, in 23 languages. Cohere states context and output limits in rounded thousands; prices were not reviewed.

No retirement announced

text inputReasoning / thinkingSelf-hostingMultilingual

Official source

Cohere

Command A Vision

command-a-vision-07-2025

Active

Context: 128K
Max output: 8,000
Distribution: API and downloadable weights
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Cohere's first image-input model; it officially supports English, Portuguese, Italian, French, German and Spanish. Cohere states context and output limits in rounded thousands; prices were not reviewed.

No retirement announced

text inputimage inputSelf-hosting

Official source

Cohere

Command A+

command-a-plus-05-2026

Active

Context: 128K
Max output: 64K
Distribution: API and downloadable weights
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Cohere's first mixture-of-experts model, combining vision input, agentic, reasoning and translation capabilities. Cohere states context and output limits in rounded thousands; prices were not reviewed.

No retirement announced

text inputimage inputReasoning / thinkingSelf-hostingMultilingual

Official source

Z.ai

GLM-5.3

glm-5.3

Active

Context: 1M
Max output: 128K
Distribution: API and downloadable weights
Input / output price: $1.40 / $4.40
Released: Not published
Knowledge cutoff: Not published

Cached-input storage is free for a limited time.

Text input only. Reasoning is always on (low, high or max). Z.ai compares it with other open-source models and Mistral's catalog lists it as an open-weight model; the license was not reviewed.

No retirement announced

text inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Z.ai

GLM-5.3-Flash

glm-5.3-flash

Active

Context: 1M
Max output: 128K
Distribution: API and downloadable weights
Input / output price: $0.15 / $0.50
Released: Not published
Knowledge cutoff: Not published

Cached-input storage is free for a limited time.

Z.ai's first native multimodal GLM-5 model, described as open-source; it also accepts files. Thinking cannot be turned off. The license was not reviewed.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Z.ai

GLM-5.3-FlashX

glm-5.3-flashx

Active

Context: 1M
Max output: 128K
Distribution: Hosted API
Input / output price: $0.37 / $1.25
Released: Not published
Knowledge cutoff: Not published

Cached-input storage is free for a limited time.

The faster serving tier of GLM-5.3-Flash (Z.ai cites about 200 tokens per second).

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.2

gpt-5.2

Active

Context: 400K
Max output: 128K
Distribution: Hosted API
Input / output price: $1.75 / $14
Released: Not published
Knowledge cutoff: August 31, 2025

Rates as listed on October 25, 2026; the provider does not say when they took effect.

OpenAI calls GPT-5.2 its previous flagship and recommends GPT-6 Astra; the gpt-5.2 model ID has no shutdown date on the deprecations page.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.4

gpt-5.4

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $2.50 / $15
Released: Not published
Knowledge cutoff: August 31, 2025

Higher rates above 272K input tokens ($5 input / $22.50 output).

Positioned by OpenAI as the affordable coding and professional-work model. No published launch date; the dated snapshot id suggests March 5, 2026.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.4 Pro

gpt-5.4-pro

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $30 / $180
Released: Not published
Knowledge cutoff: August 31, 2025

Higher rates above 272K input tokens ($60 input / $270 output).

No cached-input discount. OpenAI's pricing table prints Batch rates ($15 / $90), but the model page lists the Batch endpoint as not supported, so no Batch rate is recorded. Regional processing endpoints add 10%. Rates as listed on October 25, 2026; the provider does not say when they took effect.

Responses API only, and some requests may take several minutes.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingBatch processing

Official source

OpenAI

GPT-5.4-mini

gpt-5.4-mini

Active

Context: 400K
Max output: 128K
Distribution: Hosted API
Input / output price: $0.75 / $4.50
Released: Not published
Knowledge cutoff: August 31, 2025

No published launch date; the dated snapshot id suggests March 17, 2026.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.5

gpt-5.5

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $5 / $30
Released: Not published
Knowledge cutoff: December 1, 2025

Higher rates above 272K input tokens ($10 input / $45 output).

Still sold and priced, but no longer shown on OpenAI's models index. OpenAI has not published a launch date; the dated snapshot id suggests April 23, 2026.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

OpenAI

GPT-5.5 Pro

gpt-5.5-pro

Active

Context: 1.05M
Max output: 128K
Distribution: Hosted API
Input / output price: $30 / $180
Released: Not published
Knowledge cutoff: December 1, 2025

Higher rates above 272K input tokens ($60 input / $270 output).

No cached-input discount. Regional processing endpoints add 10%.

Available through the Responses API and Batch only, not Chat Completions. OpenAI says some requests may take several minutes and recommends background mode.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.20

grok-4.20-0309-reasoning

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.25 / $2.50
Released: Not published
Knowledge cutoff: Not published

Higher rates at or above 200K input tokens ($2.50 input / $5 output).

Batch rates are computed from xAI's stated 20% Batch API discount, which applies to all token types including cached tokens. Rates as listed on October 25, 2026; the provider does not say when they took effect.

xAI publishes no release date, maximum output limit or retirement date. The model page lists more beta and experimental aliases; use the explicit model ID.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.20 (Non-Reasoning)

grok-4.20-0309-non-reasoning

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.25 / $2.50
Released: Not published
Knowledge cutoff: Not published

Higher rates at or above 200K input tokens ($2.50 input / $5 output).

The non-reasoning variant (the model page lists Reasoning: No). xAI publishes no release date, maximum output limit or retirement date.

No retirement announced

text inputimage inputTool use / function callingStructured outputsPrompt / context caching

Official source

xAI

Grok 4.20 Multi-Agent Beta

grok-4.20-multi-agent-0309

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.25 / $2.50
Released: Not published
Knowledge cutoff: Not published

Higher rates at or above 200K input tokens ($2.50 input / $5 output).

Beta: multiple agents collaborate in parallel on deep research tasks. Lower rate limits than the single-agent model (9 requests per second). xAI publishes no release date, maximum output limit or retirement date.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

xAI

Grok 4.3

grok-4.3

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.25 / $2.50
Released: Not published
Knowledge cutoff: Not published

Higher rates at or above 200K input tokens ($2.50 input / $5 output).

The official model page lists no maximum output limit or committed retirement date. Alias mappings can change; use the explicit model ID.

No retirement announced

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Moonshot AI

Kimi K3

kimi-k3

Active

Context: 1.04M
Max output: 1.04M
Distribution: API and downloadable weights
Input / output price: $3 / $15
Released: Not published
Knowledge cutoff: Not published

Flat pricing with no long-context tier. Caching is automatic; cache writes bill once at the 5-minute or 1-hour rate.

Thinking is always on (reasoning_effort low, high or max; default max). Output defaults to 131,072 tokens and can be raised to 1,048,576. Moonshot now calls K3 its previous-generation flagship; Kimi K3.1 Pro is listed on the same pricing page. The full weights are released; the license was not reviewed.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Xiaomi

MiMo-V2.6-Flash

mimo-v2.6-flash

Active

Context: 1.04M
Max output: 128K
Distribution: API and downloadable weights
Input / output price: $0.14 / $0.28
Released: Not published
Knowledge cutoff: Not published

Xiaomi MiMo API overseas pay-as-you-go rates (domestic rates are billed in RMB). Cache writes are free for a limited time, and web search costs $5 per 1,000 calls.

Open weights on Hugging Face as XiaomiMiMo/MiMo-V2.6-Flash-RL (with a MOPD checkpoint alongside), and served as mimo-v2.6-flash on Xiaomi's MiMo API, which lists a 1M context window (1,048,576 tokens in the weights' config) and 128K maximum output.

No retirement announced

text inputimage inputvideo inputaudio inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Xiaomi

MiMo-V2.6-Pro

mimo-v2.6-pro

Active

Context: 1.04M
Max output: 128K
Distribution: API and downloadable weights
Input / output price: $0.43 / $0.87
Released: Not published
Knowledge cutoff: Not published

Xiaomi MiMo API overseas pay-as-you-go rates (domestic rates are billed in RMB). Cache writes are free for a limited time, and web search costs $5 per 1,000 calls.

Open weights on Hugging Face as XiaomiMiMo/MiMo-V2.6-Pro-RL (with a MOPD checkpoint alongside), and served as mimo-v2.6-pro on Xiaomi's MiMo API, which lists a 1M context window (1,048,576 tokens in the weights' config) and 128K maximum output.

No retirement announced

text inputimage inputvideo inputaudio inputReasoning / thinkingTool use / function callingStructured outputs

Official source

MiniMax

MiniMax M3

MiniMax-M3

Active

Context: 1M
Max output: Not published
Distribution: API and downloadable weights
Input / output price: $0.30 / $1.20
Released: Not published
Knowledge cutoff: Not published

Higher rates above 512K input tokens ($0.60 input / $2.40 output).

MiniMax shows these as a permanent 50% discount on list prices of $0.60 input and $2.40 output. Priority admission costs 1.5x.

Thinking is off by default and can be turned on. MiniMax publishes no maximum output limit; its self-hosting guide calls the open-weight deployment experimental.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingPrompt / context caching

Official source

MiniMax

MiniMax M3.1 Flash

MiniMax-M3.1-Flash

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $0.15 / $0.60
Released: Not published
Knowledge cutoff: Not published

Rates as listed on October 25, 2026; the provider does not say when they took effect.

Always thinks (thinking depth low to max, default max). MiniMax publishes no maximum output limit.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingPrompt / context caching

Official source

Meta

Muse Spark 1.3 (Contributor tier)

muse-spark-1.3-contributor

Active

Context: 1.04M
Max output: Not published
Distribution: Hosted API
Input / output price: $0.10 / $0.20
Released: Not published
Knowledge cutoff: Not published

Contributor tier rates as observed on October 8, 2026 (Meta's pricing page does not date them): discounted in exchange for permission to use prompts and completions to train future Meta models. The same model on the Standard tier (muse-spark-1.3) costs $1.25/$0.15/$4.25. Built-in web search grounding adds $2.50 per 1,000 queries.

Training-eligible tier of Muse Spark 1.3: Meta may use your prompts and completions to train future Meta models, the max reasoning level is not available, and the rate limit is 100 requests and 3,000,000 tokens per minute. Meta does not date this tier's launch, and it says audio understanding in Muse Spark 1.3 is not fully supported.

No retirement announced

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Alibaba (Qwen)

Qwen3.7-Plus

qwen3.7-plus

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Alibaba recommends it for balanced performance and cost. Model Studio states a 1M-token context window; the exact input and output limits and the prices were not reviewed.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Alibaba (Qwen)

Qwen3.8-Flash

qwen3.8-flash

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Alibaba's lower-cost option; Model Studio lists it under text generation, not image and video understanding. Model Studio states a 1M-token context window; the exact input and output limits and the prices were not reviewed.

No retirement announced

text inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Alibaba (Qwen)

Qwen3.8-Max

qwen3.8-max

Active

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: Varies by host
Released: Not published
Knowledge cutoff: Not published

Alibaba recommends it for the strongest reasoning. Model Studio states a 1M-token context window; the exact input and output limits and the prices were not reviewed.

No retirement announced

text inputimage inputvideo inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Mistral AI

Mistral Large 4

mistral-large-4

Preview

Context: 1M
Max output: Not published
Distribution: Hosted API
Input / output price: $1.36 / $4.18
Released: Oct 6, 2026
Knowledge cutoff: Not published

Standard (original) Mistral API rates; batch cached input is $0.07. Mistral's changelog announces launch pricing of 50% off for 2 weeks, shown on the pricing page as a temporary sale ($0.68 input, $0.07 cached input, $2.09 output; batch $0.34/$0.035/$1.045). Mistral publishes no exact end date for the sale, so planning uses the standard rates.

Public Preview: Mistral allows silent updates and does not guarantee general availability, and the model card lists the weights and license as coming soon (Mistral says weights ship by the end of October 2026). Mistral publishes no maximum output limit, and mistral-large-latest still points to Mistral Large 3.

No retirement date published

text inputimage inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Google

Gemini 3.1 Pro Preview

gemini-3.1-pro-preview

Preview

Context: 1.04M
Max output: 65.5K
Distribution: Hosted API
Input / output price: $2 / $12
Released: Feb 19, 2026
Knowledge cutoff: Not published

Higher rates above 200K input tokens ($4 input / $18 output).

Paid-tier price for prompts up to 200K tokens. Longer prompts use higher input and output rates.

Preview models can change or retire on shorter notice. gemini-pro-latest is a rolling alias, not a pinned release.

No retirement date published

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Meta

Muse Wonder

muse-wonder-1.0

Preview

Context: 1.04M
Max output: Not published
Distribution: Hosted API
Input / output price: $4 / $20
Released: Not published
Knowledge cutoff: Not published

Standard mode; Contributor mode is not available for Muse Wonder on Model API. No long-context premium. Web search grounding adds $2.50 per 1,000 queries.

In preview. Muse Wonder Full, the higher-capability variant, is available only in Muse Code with a subscription. Meta publishes no release date or maximum output limit, and the model is not available in South Korea.

No retirement date published

text inputimage inputvideo inputaudio inputpdf inputReasoning / thinkingTool use / function callingStructured outputs

Official source

Specifications narrow a shortlist; they do not rank quality.

The one index score shown is third-party and indicative. Task-specific benchmarks, latency, rate limits, regional availability, safety behavior, and real workload quality still require testing. For workload-specific token costs, use the API cost and context planner.

Method and limits

Facts are versioned; aliases and prices can move

The catalog is reviewed field by field; the latest review was October 25, 2026. Untouched fields retain their earlier checked dates. It preserves model IDs, effective dates, caveats, and source links rather than silently treating a rolling alias as a permanent checkpoint.

Prices shown are standard base text-token rates where the provider publishes one. Long-context tiers, regions, caching, batches, media, tools, fine-tuning, and hosted open-model prices may differ. Confirm the linked source before committing a budget or migration.

Editorial comparisons

The finder compares published specifications. These wiki articles add the qualitative context: strengths, pricing history, and deployment tradeoffs.

Questions about comparing AI models

What does the AI model comparison include?

The finder compares first-party published facts: lifecycle state, model or deployment ID, context window, maximum output, input and output modalities, selected capabilities, standard text-token pricing, distribution, parameter counts, and licenses where applicable.

Does the cheapest model always cost less for a real workload?

No. Token counts, prompt caching, batch discounts, long-context tiers, tool calls, images, audio, reasoning tokens, retries, and provider-specific fees can change the bill. Use the linked cost planner with a realistic workload.

Does open-weight mean open-source?

Not necessarily. Open-weight means downloadable model weights are available. Each model still has its own license, and some community licenses impose conditions that differ from standard open-source licenses.

Which model is best?

There is no universal winner. This tool narrows models by documented constraints; it does not invent one composite quality score. Test the finalists on representative prompts and measure quality, latency, reliability, and total cost.

Benchmark SWE-bench MMLU ARC-AGI Humanity's Last Exam Chatbot Arena Artificial Analysis Open-source AI Context window Large language model

Narrow the catalog

Compare selected models

Matching models

Facts are versioned; aliases and prices can move

Editorial comparisons

Questions about comparing AI models

Related AI Wiki articles