AnthropicReasoning modelClaude Opus 4.1
Best configuration ranks #126 of 610 on the BenchLeader Index at 58.1 ±2.7 (thinking reasoning effort). Last measured 2 Sept 2026. Released 5 Aug 2025.
- Blended price
- $30.00/M
- $15.00 in · $75.00 out
- Output speed
- 10 tok/s
- First answer
- 2.68 s
- Context
- 200k
- Overall index58
- Reasoning65
- Coding53
- Agents & tools63
- Maths59
- Knowledge58
- Instruction following53
- Human preference64
- Multimodal57
- Long context65
- Composite58
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Default” means the publisher did not say which setting was used.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| thinkingbest | 58.1 | #126 | 10 tok/s | 2.68 s | $0.029 | Agents & tools 63 · Coding 53 · Composite 58 · Human preference 64 · Instruction following 53 · Knowledge 58 · Long context 65 · Maths 59 · Multimodal 57 · Reasoning 65 |
| default | 52.6 | #242 | 10 tok/s | 2.68 s | $0.029 | Agents & tools 54 · Coding 49 · Composite 52 · Human preference 64 · Knowledge 49 · Maths 43 · Multimodal 58 · Reasoning 54 |
Benchmark results
One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Reasoning
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| GPQA Diamond | – | 77.3%#124 | Epoch AI Benchmarking Hub | |
| Humanity's Last Exam | – | 11.5%#25 | Scale AI / CAIS | |
| SimpleBench | – | 60.0%#31 | SimpleBench | |
| LMArena Hard Prompts | 1480#54 | 1477#57 | LMArena | |
| GPQA Diamond (AA)not in index | 80.9%#183 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 12.5%#216 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | 76.3%#76 | 70.0%#91 | Vals AI |
Coding
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | 73.3%#20 | Epoch AI Benchmarking Hub | |
| WeirdML | – | 45.9%#79 | WeirdML | |
| LMArena Coding | 1512#44 | 1506#54 | LMArena | |
| LMArena WebDev | – | 1389#75 | LMArena | |
| LiveCodeBench | 66.5%#95 | 64.6%#97 | Vals AI | |
| IOI | – | 12.5%#31 | Vals AI |
Agents & tools
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| Terminal-Bench | – | 38.0%#37 | Terminal-Bench | |
| GDPval | – | 43.6%#3 | OpenAI | |
| Cybench | – | 42.0%#5 | Cybench | |
| Terminal-Bench Hard | 34.3%#86 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 71.4%#147 | – | Artificial Analysis |
Maths
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | 12.6%#88 | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | 2.4%#53 | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | – | 68.9%#126 | Epoch AI Benchmarking Hub | |
| AIME (Vals) | 78.2%#48 | 44.2%#63 | Vals AI | |
| MGSM | 94.4%#3 | 94.2%#5 | Vals AI |
Knowledge
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| MMLU-Pro | 87.9%#23 | 87.2%#33 | Vals AI | |
| LegalBench | – | 83.5%#49 | Vals AI | |
| TaxEval | 73.7%#46 | 71.5%#76 | Vals AI | |
| MedQA | 93.6%#22 | 92.5%#29 | Vals AI | |
| PRBench Finance | – | 35.1%#29 | Scale AI SEAL | |
| PRBench Legal | – | 34.0%#31 | Scale AI SEAL | |
| MultiNRC | 38.4%#19 | 29.7%#23 | Scale AI SEAL |
Instruction following
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| IFBench | 55.4%#143 | – | Artificial Analysis | |
| MultiChallenge | 57.2%#14 | – | Scale AI SEAL | |
| TutorBench | 50.8%#17 | – | Scale AI SEAL |
Human preference
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| LMArena Text | 1450#66 | 1448#71 | LMArena |
Multimodal
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| MMMU-Pro | 67.9%#133 | – | Artificial Analysis | |
| VISTA | 48.4%#15 | 45.3%#28 | Scale AI SEAL | |
| MMMU (validation) | – | 77.1%#12 | MMMU |
Long context
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| AA-LCR | 76.0%#102 | – | Artificial Analysis |
Composite
| Benchmark | thinking | default | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | 144.1#76 | Epoch AI Benchmarking Hub | |
| AA Intelligence Index | 22.9#139 | 18.6#189 | Artificial Analysis |
Where it wins
Benchmarks where this configuration ranks in the top five of every configuration measured.
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Amazon Bedrock | 17 tok/s | 3.62 s | $15.00 | $75.00 | 200k | – |
| Google Vertex (Global) | 3 tok/s | 1.74 s | $15.00 | $75.00 | 200k | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $1.50 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.029 | $0.025 | 32.7 s |
| Summarise a 30-page report | 12,000 / 600 | $0.225 | $0.103 | 1.0 min |
| Code edit | 6,000 / 1,500 | $0.203 | $0.142 | 2.5 min |
| Agentic coding session | 60,000 / 4,000 | $1.20 | $0.593 | 6.7 min |
| Structured extraction | 2,000 / 200 | $0.045 | $0.025 | 22.7 s |
See also
Data as of 9 Sept 2026. Compare these configurations.