AnthropicReasoning modelClaude Opus 4.5
Best configuration ranks #83 of 610 on the BenchLeader Index at 60.6 ±2.6 (thinking reasoning effort). Last measured 1 Sept 2026. Released 24 Nov 2025.
- Blended price
- $10.00/M
- $5.00 in · $25.00 out
- Output speed
- 46 tok/s
- First answer
- 18 s
- first token 1.34 s
- Context
- 200k
- Overall index61
- Reasoning59
- Coding56
- Agents & tools74
- Maths64
- Knowledge61
- Instruction following55
- Multimodal58
- Long context65
- Composite66
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 | – | – | 45 tok/s | 1.34 s | $0.0095 | Reasoning 39 |
| medium | – | – | 45 tok/s | 1.34 s | $0.0095 | Coding 65 |
| high | 56.6 | #154 | 45 tok/s | 1.34 s | $0.0095 | Agents & tools 54 · Coding 63 · Composite 44 · Human preference 67 · Reasoning 68 |
| thinkingbest | 60.6 | #83 | 46 tok/s | 18 s | $0.0095 | Agents & tools 74 · Coding 56 · Composite 66 · Instruction following 55 · Knowledge 61 · Long context 65 · Maths 64 · Multimodal 58 · Reasoning 59 |
| default | 58.0 | #132 | 45 tok/s | 1.34 s | $0.0095 | Agents & tools 68 · Coding 57 · Composite 59 · Human preference 67 · Instruction following 45 · Knowledge 57 · Long context 62 · Maths 51 · Multimodal 58 · Reasoning 62 |
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 | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | – | – | 86.0%#71 | Epoch AI Benchmarking Hub | |
| Humanity's Last Exam | – | – | – | – | 25.2%#13 | Scale AI / CAIS | |
| SimpleBench | – | – | – | – | 62.0%#22 | SimpleBench | |
| LMArena Hard Prompts | – | – | 1499#25 | – | 1498#28 | LMArena | |
| LiveBench Reasoningnot in index | – | – | 80.1%#38 | – | – | LiveBench | |
| GPQA Diamond (AA)not in index | – | – | – | 86.6%#109 | 81.0%#180 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | – | – | – | 30.1%#102 | 13.2%#210 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | 85.9%#43 | 79.5%#68 | Vals AI | |
| Kagi LLM Benchmark | – | – | – | 80.2%#8 | 70.7%#25 | Kagi LLM Benchmark | |
| ARC-AGI-1 | 40.0%#132 | – | – | 80.0%#81 | – | ARC Prize | |
| ARC-AGI-2 | 7.8%#110 | – | – | 37.6%#81 | – | ARC Prize |
Coding
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch) | – | – | – | – | 76.7%#9 | Epoch AI Benchmarking Hub | |
| WeirdML | – | – | – | – | 63.7%#38 | WeirdML | |
| GSO-Bench | – | – | – | – | 26.5%#10 | GSO-Bench | |
| LMArena Coding | – | – | 1531#13 | – | 1523#21 | LMArena | |
| LMArena WebDev | – | – | 1495#41 | – | 1468#48 | LMArena | |
| LiveBench Codingnot in index | – | – | 79.7%#14 | – | – | LiveBench | |
| LiveCodeBench | – | – | – | 83.7%#45 | 75.0%#78 | Vals AI | |
| IOI | – | – | – | 20.3%#23 | 23.6%#19 | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | 76.4%#35 | – | Vals AI | |
| SWE-Bench Pro | – | – | – | – | 45.9%#6 | Scale AI SEAL | |
| SWE-bench Verified (bash only) | – | 74.4%#5 | 76.8%#1 | – | – | SWE-bench | |
| SWE-bench Verified (any scaffold)not in index | – | – | – | – | 79.2%#1 | SWE-bench |
Agents & tools
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| Terminal-Bench | – | – | – | – | 63.1%#12 | Terminal-Bench | |
| GDPval | – | – | – | – | 45.5%#2 | OpenAI | |
| Cybench | – | – | – | – | 82.0%#2 | Cybench | |
| Remote Labor Index | – | – | – | – | 3.8%#6 | Scale AI / CAIS | |
| APEX-Agents | – | – | – | – | 20.7%#35 | Mercor | |
| LiveBench Agentic Codingnot in index | – | – | 39.7%#49 | – | – | LiveBench | |
| Terminal-Bench Hard | – | – | – | 47.0%#27 | 40.9%#51 | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | – | – | – | 89.5%#68 | 86.3%#87 | Artificial Analysis | |
| MCP Atlas | – | – | 69.8%#19 | – | – | Scale AI SEAL | |
| BFCL Overall | – | – | – | – | 77.5%#1 | Berkeley Function Calling Leaderboard | |
| τ²-bench | – | – | 85.3%#2 | – | – | τ²-bench |
Maths
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | 34.4%#53 | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | 4.9%#51 | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | – | – | – | – | 86.1%#79 | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | 36.0%#25 | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | 90.4%#23 | – | – | LiveBench | |
| AIME (Vals) | – | – | – | 95.4%#12 | 76.9%#50 | Vals AI | |
| MGSM | – | – | – | 95.2%#1 | 94.8%#2 | Vals AI |
Knowledge
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | 45.7%#31 | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | 74.4%#28 | – | – | LiveBench | |
| AA-Omniscience | – | – | – | 14#56 | -4.1#112 | Artificial Analysis | |
| MMLU-Pro | – | – | – | 87.3%#31 | 85.6%#55 | Vals AI | |
| LegalBench | – | – | – | 84.6%#28 | 82.8%#57 | Vals AI | |
| CorpFin | – | – | – | 65.1%#34 | 61.3%#54 | Vals AI | |
| TaxEval | – | – | – | 74.9%#25 | 74.3%#37 | Vals AI | |
| MedQA | – | – | – | 95.9%#10 | 93.2%#24 | Vals AI | |
| PRBench Finance | – | – | – | 46.2%#16 | – | Scale AI SEAL | |
| PRBench Legal | – | – | – | 44.2%#16 | – | Scale AI SEAL | |
| MultiNRC | – | – | – | 48.6%#12 | 41.2%#18 | Scale AI SEAL |
Instruction following
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | 81.3%#20 | – | – | LiveBench | |
| IFBench | – | – | – | 58.0%#131 | 43.0%#225 | Artificial Analysis | |
| MultiChallenge | – | – | – | 59.0%#12 | – | Scale AI SEAL | |
| TutorBench | – | – | – | 51.2%#16 | 49.8%#18 | Scale AI SEAL |
Human preference
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | 1473#34 | – | 1469#40 | LMArena |
Multimodal
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| MMMU-Pro | – | – | – | 74.0%#88 | 71.2%#108 | Artificial Analysis | |
| VISTA | – | – | – | 46.4%#21 | 45.3%#27 | Scale AI SEAL | |
| MMMU (validation) | – | – | – | – | 80.7%#5 | MMMU | |
| MMMU-Pro (official)not in index | – | – | – | – | 73.9%#8 | MMMU |
Long context
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| AA-LCR | – | – | – | 77.3%#91 | 70.7%#154 | Artificial Analysis |
Composite
| Benchmark | no reasoning | medium | high | thinking | default | Source | Trend |
|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | 150.1#43 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | 72.6%#39 | – | – | LiveBench | |
| AA Intelligence Index | – | – | – | 29.1#85 | 23.7#128 | 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 |
|---|---|---|---|---|---|---|
| Azure | 71 tok/s | 1.28 s | $5.00 | $25.00 | 200k | – |
| Google Vertex | 56 tok/s | 1.26 s | $5.00 | $25.00 | 200k | – |
| Claude Platform on AWS | 53 tok/s | 1.67 s | $5.00 | $25.00 | 200k | – |
| Amazon Bedrock | 36 tok/s | 1.75 s | $5.00 | $25.00 | 200k | – |
| Anthropic | 13 tok/s | 1.22 s | $5.00 | $25.00 | 200k | – |
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 | 24.1 s |
| Summarise a 30-page report | 12,000 / 600 | $0.075 | $0.035 | 30.6 s |
| Code edit | 6,000 / 1,500 | $0.068 | $0.047 | 50.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.400 | $0.198 | 1.8 min |
| Structured extraction | 2,000 / 200 | $0.015 | $0.0083 | 21.9 s |
See also
Data as of 9 Sept 2026. Compare these configurations.