中文 ← MyDocs

Models overview

Claude is a family of state-of-the-art large language models developed by Anthropic. This guide introduces the available models and compares their performance.

Choosing a model

If you're unsure which model to use, consider starting with Claude Opus 4.7 for the most complex tasks. It is our most capable generally available model, with a step-change improvement in agentic coding over Claude Opus 4.6.

All current Claude models support text and image input, text output, multilingual capabilities, and vision. Models are available through the Claude API, Claude Platform on AWS, Amazon Bedrock, Vertex AI, and Microsoft Foundry.

Once you've picked a model, learn how to make your first API call.

Latest models comparison

FeatureClaude Opus 4.7Claude Sonnet 4.6Claude Haiku 4.5
DescriptionOur most capable generally available model for complex reasoning and agentic codingThe best combination of speed and intelligenceThe fastest model with near-frontier intelligence
Claude API IDclaude-opus-4-7claude-sonnet-4-6claude-haiku-4-5-20251001
Claude API aliasclaude-opus-4-7claude-sonnet-4-6claude-haiku-4-5
AWS Bedrock IDanthropic.claude-opus-4-73anthropic.claude-sonnet-4-6anthropic.claude-haiku-4-5-20251001-v1:0
Vertex AI IDclaude-opus-4-7claude-sonnet-4-6claude-haiku-4-5@20251001
Pricing1$5 / input MTok
$25 / output MTok		$3 / input MTok
$15 / output MTok		$1 / input MTok
$5 / output MTok
Extended thinkingNoYesYes
Adaptive thinkingYesYesNo
Priority TierYesYesYes
Comparative latencyModerateFastFastest
Context window
<Tooltip tooltipContent="~555k words \\ ~2.5M unicode characters (Opus 4.7 uses a new tokenizer)">
1M tokens
<Tooltip tooltipContent="~750k words \\ ~3.4M unicode characters">
1M tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
Max output128k tokens64k tokens64k tokens
Reliable knowledge cutoffJan 20262Aug 20252Feb 2025
Training data cutoffJan 2026Jan 2026Jul 2025

1 - See Pricing for complete pricing information including Batch API discounts and prompt caching rates.

2 - Reliable knowledge cutoff indicates the date through which a model's knowledge is most extensive and reliable. Training data cutoff is the broader date range of training data used. For more information, see Anthropic's Transparency Hub.

3 - Claude Opus 4.7 is available on Bedrock through Claude in Amazon Bedrock (the Messages-API Bedrock endpoint).

Info

Claude Mythos Preview is offered separately as a research preview model for defensive cybersecurity workflows as part of Project Glasswing. Access is invitation-only and there is no self-serve sign-up.

Note
Every Claude model ID is a pinned snapshot. Models with a date in the ID (for example, 20250929) are fixed to that specific release. Starting with the Claude 4.6 generation, model IDs use a dateless format that is also a pinned snapshot, not an evergreen pointer. For models before the 4.6 generation, entries in the Claude API alias column are convenience pointers that resolve to a dated model ID. For details on the naming convention and how versioning works, see Model IDs and versioning.
Note
Starting with Claude Sonnet 4.5 and all subsequent models (including Claude Sonnet 4.6), Bedrock offers two endpoint types: global endpoints (dynamic routing for maximum availability) and regional endpoints (guaranteed data routing through specific geographic regions). Vertex AI offers three endpoint types: global endpoints, multi-region endpoints (dynamic routing within a geographic area), and regional endpoints. For more information, see Cloud platform pricing.
Note
Claude Platform on AWS uses the same model IDs as the Claude API (for example, claude-opus-4-6), not Bedrock-style IDs. Model lifecycle on Claude Platform on AWS follows Anthropic's first-party Model deprecations, not Bedrock's. See Available models for the model list.
Tip

You can query model capabilities and token limits programmatically with the Models API. The response includes max_input_tokens, max_tokens, and a capabilities object for every available model.

Note

The Max output values above apply to the synchronous Messages API. On the Message Batches API, Opus 4.7, Opus 4.6, and Sonnet 4.6 support up to 300k output tokens by using the output-300k-2026-03-24 beta header.

Legacy models

The following models are still available. Consider migrating to current models for improved performance:

FeatureClaude Opus 4.6Claude Sonnet 4.5Claude Opus 4.5Claude Opus 4.1Claude Sonnet 4 (deprecated)Claude Opus 4 (deprecated)
Claude API IDclaude-opus-4-6claude-sonnet-4-5-20250929claude-opus-4-5-20251101claude-opus-4-1-20250805claude-sonnet-4-20250514claude-opus-4-20250514
Claude API aliasclaude-opus-4-6claude-sonnet-4-5claude-opus-4-5claude-opus-4-1claude-sonnet-4-0claude-opus-4-0
AWS Bedrock IDanthropic.claude-opus-4-6-v1anthropic.claude-sonnet-4-5-20250929-v1:0anthropic.claude-opus-4-5-20251101-v1:0anthropic.claude-opus-4-1-20250805-v1:0anthropic.claude-sonnet-4-20250514-v1:0anthropic.claude-opus-4-20250514-v1:0
Vertex AI IDclaude-opus-4-6claude-sonnet-4-5@20250929claude-opus-4-5@20251101claude-opus-4-1@20250805claude-sonnet-4@20250514claude-opus-4@20250514
Pricing$5 / input MTok
$25 / output MTok		$3 / input MTok
$15 / output MTok		$5 / input MTok
$25 / output MTok		$15 / input MTok
$75 / output MTok		$3 / input MTok
$15 / output MTok		$15 / input MTok
$75 / output MTok
Extended thinkingYesYesYesYesYesYes
Priority TierYesYesYesYesYesYes
Comparative latencyModerateFastModerateModerateFastModerate
Context window
<Tooltip tooltipContent="~750k words \\ ~3.4M unicode characters">
1M tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
<Tooltip tooltipContent="~150k words \\ ~680k unicode characters">
200k tokens
Max output128k tokens64k tokens64k tokens32k tokens64k tokens32k tokens
Reliable knowledge cutoffMay 20251Jan 20251May 20251Jan 20251Jan 20251Jan 20251
Training data cutoffAug 2025Jul 2025Aug 2025Mar 2025Mar 2025Mar 2025
Warning

Claude Sonnet 4 (claude-sonnet-4-20250514) and Claude Opus 4 (claude-opus-4-20250514) are deprecated and will be retired on June 15, 2026. Migrate to Claude Sonnet 4.6 and Claude Opus 4.7 respectively before the retirement date.

See model deprecations for details.

1 - Reliable knowledge cutoff indicates the date through which a model's knowledge is most extensive and reliable. Training data cutoff is the broader date range of training data used.

Prompt and output performance

Claude 4 models excel in:

  • Performance: Top-tier results in reasoning, coding, multilingual tasks, long-context handling, honesty, and image processing. See the Claude 4 blog post for more information.

  • Engaging responses: Claude models are ideal for applications that require rich, human-like interactions.

  • Output quality: When migrating from previous model generations to Claude 4, you may notice larger improvements in overall performance.

Migrating to Claude Opus 4.7

If you're currently using Claude Opus 4.6 or older Claude models, consider migrating to Claude Opus 4.7 to take advantage of improved intelligence and a step-change jump in agentic coding. For detailed migration instructions, see Migrating to Claude Opus 4.7.

Get started with Claude

If you're ready to start exploring what Claude can do for you, dive in! Whether you're a developer looking to integrate Claude into your applications or a user wanting to experience the power of AI firsthand, the following resources can help.

Note
Looking to chat with Claude? Visit claude.ai!
◆Intro to Claude

Explore Claude's capabilities and development flow.

◆Quickstart

Learn how to make your first API call in minutes.

◆Claude Console

Craft and test powerful prompts directly in your browser.

If you have any questions or need assistance, don't hesitate to reach out to the support team or consult the Discord community.