Model · Anthropic · Text
Claude Opus 5
Anthropic's default recommendation for most workloads — frontier reasoning and long-horizon agentic coding at half the price of Claude Fable 5. Succeeds Claude Opus 4.8 as the Opus-tier flagship.
- Modality
- Text
- License
- Proprietary (Proprietary)
- Context window
- 1,000,000 tokens 1M tokens is both the default and the maximum — there is no smaller context variant. Up to 128k output, or 300k on the Batch API with the output-300k beta header.
- Released
- July 24, 2026
- Last verified
- September 8, 2026
- Runs locally
- No
- Also handles
- vision
Strengths
- Deep reasoning sustained across long, multistep problem chains
- Agentic coding and long-horizon work — multi-file features and large refactors completed without leaving stubs or placeholders
- Test-time compute scaling: extra effort (up to the `max` level) converts into better results
- Frontier capability at half the per-token cost of Claude Fable 5 ($5/$25 vs $10/$50 per MTok)
Weaknesses
- Thinking is on by default — code that reads `content[0].text` must select content blocks by `type`, and thinking tokens bill as output
- Disabling thinking while setting effort to `xhigh` or `max` returns a 400 error
- Over-verifies its own work on simple tasks, costing extra tokens and latency
- Priority Tier is not supported
- Closed weights — cannot be self-hosted
Try it
| Where | Type | Notes |
|---|---|---|
| Anthropic API | hosted-api | API key required |
| AWS Bedrock | hosted-api | Enterprise / regional |
| Google Vertex AI | hosted-api | GCP integration |
| Claude.ai | product | Paid plans (Pro and above) |
Used in solutions
Version history
- Claude Fable 5.1 Sep 2026
- Claude Mythos 5.1 Sep 2026
- Claude Opus 5 Jul 2026 Current
- Claude Sonnet 5 Jun 2026
- Claude Fable 5 Jun 2026 Deprecated
Official sources
- Model docs docs
- What's new in Opus 5 docs
- Models overview docs
Change log
- — Initial entry. Released 2026-07-24 as the Opus-tier flagship, superseding Claude Opus 4.8. 1M-token context (default and maximum), $5/$25 per MTok, adaptive thinking on by default with a low/medium/high/xhigh/max effort ladder.
Esc