Claude Opus 4.8
Top Claude Opus tier for the hardest reasoning, coding, and long-horizon agents.
- Capability
- 159.9ECI · #22 of 158
- Input
- $5.00per 1M tokens
- Output
- $25.00per 1M tokens
- Context
- 1M128K max output
What it costs
- Input
- $5.00per million tokens
- Output
- $25.00per million tokens
- Cached input
- $0.50per million tokens
- Cache write
- $6.25per million tokens
- Blended (3:1)
- $10.00pricier than 94% of priced models
Typical monthly bills
| Monthly usage | Estimated cost |
|---|---|
| Side project2M input + 0.5M output tokens | $22.50 |
| Team assistant25M input + 5M output tokens | $250.00 |
| Production app250M input + 50M output tokens | $2,500 |
Official Anthropic API price, as listed on models.dev.
Independent benchmarks
#22 of 158 scored models
The shaded band is Epoch AI’s confidence range (157.8–162.2); the tick marks the median scored model.
Scores from Epoch AI, run independently of Anthropic.
The details
- Lab
- Anthropic
- Released
- May 28, 2026
- Knowledge cutoff
- Jan 2026
- Context window
- 1,000,000 tokens
- Max output
- 128,000 tokens
- Inputs
- Text, Images, PDFs
- Output
- Text
- Reasoning
- Adjustable effort low · medium · high · xhigh · max
- Tool calling
- Yes
- Structured output
- Yes
- Weights
- Proprietary
- API model ID
claude-opus-4-8on Anthropic- Availability
- 57 API providerslisted on models.dev
What Anthropic claims
Published by the lab at launch. Settings vary, so compare these only with care.
| Benchmark | Score | Setting | Source |
|---|---|---|---|
| SWE-Bench Pro | 69.2 resolve rate | — | Source |
| Terminal-Bench v2.1 | 74.6 success rate | — | Source |
| SWE-Bench Verified | 88.6 resolved | — | Source |
| Humanity's Last Exam | 49.8 | no tools | Source |
| Humanity's Last Exam | 57.9 | with tools | Source |
| OSWorld-Verified | 83.4 success rate | — | Source |
| FrontierCode | 13.4 | high effort | Source |
More from Anthropic
About Claude Opus 4.8
How much does Claude Opus 4.8 cost?
Claude Opus 4.8 costs $5.00 input / $25.00 output per million tokens (official Anthropic API price). Cached input is $0.50 per million tokens, and writing to the cache costs $6.25. At a 3:1 input-to-output mix that is $10.00 per million tokens, more expensive than 94% of the 469 priced models we track.
What is the context window of Claude Opus 4.8?
Claude Opus 4.8 accepts up to 1,000,000 tokens per request and can write up to 128,000 tokens in one response.
How good is Claude Opus 4.8?
Epoch AI gives Claude Opus 4.8 a Capabilities Index score of 159.9 (likely range 157.8–162.2), ranking it #22 of 158 models Epoch has scored. Epoch AI benchmark results: GPQA Diamond 91.0%, FrontierMath Tiers 1–3 80.0%, OTIS Mock AIME 2024–2025 98.3%, SimpleQA Verified 53.0%.
Is Claude Opus 4.8 open source?
No. Claude Opus 4.8 is proprietary; you use it through Anthropic’s API or partner platforms.
What inputs does Claude Opus 4.8 support?
Claude Opus 4.8 accepts text, images and PDFs and replies in text. It is a reasoning model with low, medium, high, xhigh and max effort settings, supports tool calling and can return structured JSON output.
When was Claude Opus 4.8 released?
Anthropic released Claude Opus 4.8 on May 28, 2026. Its training data runs to Jan 2026.
What are the best alternatives to Claude Opus 4.8?
The closest current models from other labs on capability, price and release date are GPT-5.5 (OpenAI, ECI 160.2, $5.00 / $30.00), Kimi K3 (Moonshot AI, ECI 158.6, $3.00 / $15.00), Muse Wonder (Meta, ECI 161.7, $4.00 / $20.00) and Gemini 3.1 Pro Preview (Google, ECI 156.3, $2.00 / $12.00).