Claude Opus 4.6
AnthropicProprietaryNon-reasoningVerifiedReleased Feb 1, 2026Input: text, image, file
Overall score
62.3
Rank 48 of 222 · 50.1–74.5
Context window
1M
1,000,000 tokens
Price per 1M tokens
$5.00 in · $25.00 out
✓ Verified on Anthropic's pricing page
Blended price
$10.00
3:1 input:output tokens
Output speed
—
Not measured yet
Category scores
Scores per capability area with the model's rank among ranked models in that category.
- 1Knowledge#5281.9
- 2Multimodal#2574.5
- 3Agentic & tool use#5969.8
- 4Math69.1
- 5Coding55.0
- 6Instruction following#8752.4
- 7Reasoning51.2
Benchmark results (36)
As reported in system cards, papers and public evaluations. Position compares this model with every other model reporting the same benchmark.
Agentic & tool use
| Gert Labs | 61.85 | #10 of 52 |
| BrowseComp | 83.7 | #16 of 50 |
| Terminal-Bench 2.0 | 65.4 | #12 of 39 |
| Claw-Eval | 70.4 | #8 of 36 |
| OSWorld-Verified | 72.7 | #20 of 35 |
| JobBench | 36.7 | #19 of 34 |
| CyberGym | 66.6 | #20 of 25 |
| DeepSearchQA | 73.7 | #15 of 20 |
| ResearchClawBench | 19.9 | #4 of 19 |
| ApprenticeBench | 5 | #17 of 18 |
| AndroidWorld | 62.0 | #8 of 8 |
Knowledge
| GPQA | 91.3 | #18 of 117 |
| HLE | 53.0 | #13 of 69 |
| MMLU-Pro | 82.0 | #30 of 67 |
| HLE w/o tools | 40.0 | #23 of 43 |
| SuperGPQA | 95.0 | #1 of 20 |
| HealthBench Hard | 14.8 | #12 of 12 |
| MMLU-Pro (Arcee) | 89.1 | #1 of 5 |
| MedXpertQA (Text) | 52.1 | #4 of 5 |
Coding
| SWE-bench Verified | 80.8 | #13 of 98 |
| SWE-bench Pro | 53.4 | #53 of 69 |
| FrontierCode 1.1 Main | 26.9 | #15 of 16 |
| React Native Evals | 84.1 | #4 of 15 |
| SWE-Rebench | 65.3 | #1 of 13 |
| LiveCodeBench Pro | 70.7 | #5 of 6 |
| SWE-bench Verified* | 75.6 | #1 of 5 |
Multimodal
| Design Arena Website | 1291 | #18 of 91 |
| MMMU-Pro | 77.3 | #25 of 48 |
| ParseBench | 54.07 | #31 of 37 |
| ERQA | 51.6 | #15 of 16 |
| ScreenSpot Pro | 83.1 | #7 of 15 |
| MedXpertQA (MM) | 64.8 | #7 of 8 |
Math
| FrontierMath v2 (Tiers 1-3) | 65.965 | #28 of 64 |
| FrontierMath v2 (Tier 4) | 26.829 | #30 of 51 |
| AIME25 (Arcee) | 99.8 | #1 of 5 |
Reasoning
| ARC-AGI-3 | 0.5 | #12 of 22 |
Pricing by source
USD per 1M tokens.
| Source | Input | Output | Cached in |
|---|---|---|---|
| ✓ Official (Anthropic) checked Oct 17, 2026 · Claude Opus 4.6 | $5.00 | $25.00 | $0.50 |
| BenchLM list price | $5.00 | $25.00 | $0.50 |
| OpenRouter anthropic/claude-opus-4.6 | $5.00 | $25.00 | $0.50 |
Anthropic public pricing updated for the current Opus tier; Anthropic's April 16, 2026 Claude Opus 4.7 announcement says pricing remains the same as Opus 4.6 at $5/$25 per million tokens.
Arena ratings
Human-preference ratings from LMArena. Best-rated variant shown.
| Text(High) | 1502 | #6 · 81.4K votes |
| WebDev (coding)(High) | 1547 | #44 · 19.4K votes |
| Vision(High) | 1316 | #6 · 22.3K votes |
| Documents(high) | 1507 | #3 · 27.9K votes |
Speed
Measured by us through OpenRouter (median of trials).
Not measured yet.
Related models
Similar score
- Ornith-1.5-397B62.3 · —
- Qwen3.7 Max62.6 · $2.21
- Gemini 3.5 Flash61.9 · $3.38
- Muse Spark Open62.8 · $2.00
- Gemini 3.6 Flash62.8 · $1.50