AnthropicClaude Opus 4.7
Best configuration ranks #25 of 610 on the BenchLeader Index at 65.5 ±2.5. Last measured 8 Sept 2026. Released 16 Apr 2026.
- Blended price
- $10.00/M
- $5.00 in · $25.00 out
- Output speed
- 46 tok/s
- First answer
- 21 s
- first token 1.55 s
- Context
- 1M
- Overall index66
- Reasoning66
- Coding65
- Agents & tools69
- Maths65
- Knowledge66
- Instruction following59
- Human preference69
- Multimodal66
- Long context66
- Composite80
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 |
|---|---|---|---|---|---|---|
| no reasoning | 60.1 | #94 | 44 tok/s | 1.57 s | $0.0095 | Agents & tools 81 · Composite 68 · Instruction following 46 · Knowledge 71 · Long context 64 · Multimodal 61 |
| high | 61.3 | #71 | 46 tok/s | 21 s | $0.0095 | Coding 68 · Human preference 71 · Multimodal 68 · Reasoning 55 |
| xhigh | 57.2 | #146 | 46 tok/s | 21 s | $0.0095 | Composite 56 · Knowledge 60 · Maths 60 · Reasoning 65 |
| max | 59.3 | #107 | 46 tok/s | 21 s | $0.0095 | Agents & tools 57 · Coding 67 · Maths 61 · Reasoning 63 |
| thinking | – | – | 46 tok/s | 21 s | $0.0095 | Reasoning 73 |
| defaultbest | 65.5 | #25 | 46 tok/s | 21 s | $0.0095 | Agents & tools 69 · Coding 65 · Composite 80 · Human preference 69 · Instruction following 59 · Knowledge 66 · Long context 66 · Maths 65 · Multimodal 66 · Reasoning 66 |
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 | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | 90.2%#37 | 86.4%#66 | – | – | Epoch AI Benchmarking Hub | |
| Humanity's Last Exam | – | – | – | – | – | 36.2%#7 | Scale AI / CAIS | |
| SimpleBench | – | – | – | – | – | 61.7%#23 | SimpleBench | |
| LMArena Hard Prompts | – | 1526#4 | – | – | – | 1519#6 | LMArena | |
| LiveBench Reasoningnot in index | – | – | 87.2%#21 | – | – | – | LiveBench | |
| GPQA Diamond (AA)not in index | 88.5%#88 | – | – | – | – | 91.4%#47 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 33.3%#92 | – | – | – | – | 42.3%#41 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | 90.2%#25 | Vals AI | |
| Kagi LLM Benchmark | – | – | – | – | 80.7%#7 | 73.3%#18 | Kagi LLM Benchmark | |
| ARC-AGI-3 | – | 0.2%#30 | – | – | – | – | ARC Prize |
Coding
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | – | – | 83.5%#1 | – | – | Epoch AI Benchmarking Hub | |
| SciCode | – | – | – | 54.5%#29 | – | – | SciCode | |
| WeirdML | – | 76.4%#22 | – | 75.5%#25 | – | 76.4%#23 | WeirdML | |
| FrontierCode | – | – | – | – | – | 38.5%#16 | Cognition | |
| GSO-Bench | – | 44.1%#2 | – | – | – | 44.1%#2 | GSO-Bench | |
| LMArena Coding | – | 1552#1 | – | – | – | 1547#4 | LMArena | |
| LMArena WebDev | – | 1555#24 | – | – | – | 1557#22 | LMArena | |
| LiveBench Codingnot in index | – | – | 82.1%#7 | – | – | – | LiveBench | |
| LiveCodeBench | – | – | – | – | – | 85.1%#34 | Vals AI | |
| IOI | – | – | – | – | – | 47.1%#10 | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | 82.0%#21 | Vals AI |
Agents & tools
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | – | – | – | 80.2%#3 | Terminal-Bench | |
| OSWorld-Verified 2.0 | – | – | – | 18.2%#8 | – | – | OSWorld | |
| APEX-Agents | – | – | – | 33.9%#21 | – | – | Mercor | |
| LiveBench Agentic Codingnot in index | – | – | 50.7%#29 | – | – | – | LiveBench | |
| Terminal-Bench Hard | 54.5%#15 | – | – | – | – | 51.5%#21 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 74.0%#139 | – | – | – | – | 88.6%#73 | Artificial Analysis | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | 68.5%#24 | Vals AI | |
| MCP Atlas | – | – | – | 79.1%#11 | – | – | Scale AI SEAL | |
| HiL-Bench | – | – | – | – | – | 41.7%#5 | Scale AI SEAL |
Maths
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | 70.2%#17 | – | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | 31.7%#25 | – | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | – | – | 97.8%#16 | 86.7%#71 | – | – | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | 54.0%#15 | – | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | 92.8%#15 | – | – | – | LiveBench | |
| AIME (Vals) | – | – | – | – | – | 96.3%#7 | Vals AI | |
| AIME 2026 | – | – | 95.8%#13 | – | – | – | MathArena | |
| HMMT February 2026 | – | – | 93.9%#10 | – | – | – | MathArena | |
| MathArena Apex | – | – | 40.6%#7 | – | – | – | MathArena |
Knowledge
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | 51.7%#19 | – | – | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | 78.3%#17 | – | – | – | LiveBench | |
| AA-Omniscience | 14.8#53 | – | – | – | – | 27.3#25 | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | 89.9%#8 | Vals AI | |
| LegalBench | – | – | – | – | – | 85.3%#21 | Vals AI | |
| CorpFin | – | – | – | – | – | 66.1%#22 | Vals AI | |
| TaxEval | – | – | – | – | – | 75.3%#20 | Vals AI |
Instruction following
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | 77.9%#30 | – | – | – | LiveBench | |
| IFBench | 43.6%#217 | – | – | – | – | 58.6%#128 | Artificial Analysis |
Human preference
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | 1502#4 | – | – | – | 1495#7 | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | 1311#5 | EQ-Bench |
Multimodal
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | 1316#4 | – | – | – | 1317#3 | LMArena | |
| MMMU-Pro | 76.4%#63 | – | – | – | – | 78.8%#47 | Artificial Analysis |
Long context
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| AA-LCR | 75.7%#107 | – | – | – | – | 78.7%#75 | Artificial Analysis |
Composite
| Benchmark | no reasoning | high | xhigh | max | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | 156.3#16 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | 76.5%#20 | – | – | – | LiveBench | |
| AA Intelligence Index | 30.9#74 | – | – | – | – | 40.7#32 | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | 56.1#19 | Vals AI |
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 |
|---|---|---|---|---|---|---|
| Google Vertex | 69 tok/s | 1.50 s | $5.00 | $25.00 | 1M | – |
| Amazon Bedrock (Global) | 52 tok/s | 1.60 s | $5.00 | $25.00 | 1M | – |
| Anthropic | 50 tok/s | 1.70 s | $5.00 | $25.00 | 1M | – |
| Claude Platform on AWS | 43 tok/s | 1.25 s | $5.00 | $25.00 | 1M | – |
Price history
Listed price per 1M tokens over time, as recorded by OpenRouter for the provider with the longest history.
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 $0.500 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0095 | $0.0081 | 27.7 s |
| Summarise a 30-page report | 12,000 / 600 | $0.075 | $0.035 | 34.2 s |
| Code edit | 6,000 / 1,500 | $0.068 | $0.047 | 53.7 s |
| Agentic coding session | 60,000 / 4,000 | $0.400 | $0.198 | 1.8 min |
| Structured extraction | 2,000 / 200 | $0.015 | $0.0083 | 25.5 s |
See also
Data as of 9 Sept 2026. Compare these configurations.