AnthropicReasoning modelClaude Opus 5
Claude Opus 5 is an Anthropic proprietary reasoning model, released 24 Jul 2026. Its best configuration (high reasoning effort) ranks #14 of 756 on the BenchLeader Index at 68.4 ±3.7, in the top 25. The ± is the point: 59 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (85) and lowest in long context (65). At $10.00 per million tokens blended it is among the most expensive ranked models. Output speed of 54 tokens per second puts it slower than most, with a first answer in 19.8 s. It has been measured at 6 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 2 Oct 2026.
- Blended price
- $10.00/M
- $5.00 in · $25.00 out
- Output speed
- 54 tok/s
- measured by Artificial Analysis
- First answer
- 20 s
- first token 3.10 s
- Context
- 1M
- Full answer
- 29 s
- median, reasoning included
- Cost per run
- $3.61
- one full Intelligence Index run
- Released
- 24 Jul 2026
- Overall index68
- Reasoning73
- Coding69
- Agents & tools65
- Maths71
- Knowledge77
- Human preference69
- Multimodal67
- Long context65
- Composite85
Versions
Anthropic has shipped 9 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Claude Opus 5.5high | 22 Sept 2026 | 70.1 | #2 |
| Claude Opus 5highthis page | 24 Jul 2026 | 68.4 | #14 |
| Claude Opus 4.8max | 28 May 2026 | 62.4 | #82 |
| Claude Opus 4.7 | 16 Apr 2026 | 63.8 | #59 |
| Claude Opus 4.6 | 17 Dec 2025 | 63.1 | #68 |
| Claude Opus 4.5thinking | 24 Nov 2025 | 60.2 | #106 |
| Claude Opus 4.1thinking | 5 Aug 2025 | 56.7 | #181 |
| Claude Opus 4thinking | 22 May 2025 | 55.3 | #214 |
| Claude 3 Opus | 4 Mar 2024 | 39.3 | #664 |
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. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| low | 61.7 | #90 | 55 tok/s | 2.87 s | $0.0095 | Agents & tools 52 · Coding 52 · Composite 75 · Knowledge 75 · Long context 66 · Maths 65 · Multimodal 63 · Reasoning 72 |
| medium | 65.0 | #44 | 52 tok/s | 5.27 s | $0.0095 | Agents & tools 60 · Coding 56 · Composite 81 · Knowledge 76 · Long context 66 · Multimodal 65 · Reasoning 86 |
| highbest | 68.4 | #14 | 54 tok/s | 20 s | $0.0095 | Agents & tools 65 · Coding 69 · Composite 85 · Human preference 69 · Knowledge 77 · Long context 65 · Maths 71 · Multimodal 67 · Reasoning 73 |
| xhigh | 67.9 | #20 | 56 tok/s | 45 s | $0.0095 | Agents & tools 66 · Coding 62 · Composite 86 · Knowledge 78 · Long context 66 · Multimodal 67 · Reasoning 88 |
| max | 68.2 | #17 | 58 tok/s | 68 s | $0.0095 | Agents & tools 65 · Coding 71 · Composite 76 · Human preference 69 · Knowledge 72 · Long context 65 · Maths 71 · Multimodal 68 · Reasoning 75 |
| not stated | 66.3 | #33 | 67 tok/s | 3.10 s | $0.0095 | Agents & tools 71 · Coding 68 · Human preference 78 · Knowledge 66 · Maths 67 · Reasoning 71 |
What more thinking costs
Turning the effort up buys index points and multiplies the bill. Here max costs 5.3× low for +6.5 on the index — 2.7 points per doubling of spend. Compare that with other models.
How its index has moved
8 Sept 2026 to 3 Oct 2026- high
- max
- xhigh
- not stated
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
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 | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | 87.9%#59 | – | – | – | 93.9%#14 | 92.9%#21 | Epoch AI Benchmarking Hub | |
| SimpleBench | – | – | – | – | – | 80.6%#2 | SimpleBench | |
| LMArena Hard Prompts | – | – | 1515#12 | – | 1514#13 | – | LMArena | |
| LiveBench Reasoningnot in index | – | – | – | – | 91.2%#8 | – | LiveBench | |
| GPQA Diamond (AA)not in index | 88.9%#82 | 91.9%#44 | 93.7%#11 | 93.7%#11 | 93.2%#22 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 43.4%#52 | 51.3%#22 | 52.8%#18 | 54.4%#14 | 54.9%#10 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | – | – | 93.4%#7 | Vals AI | |
| ARC-AGI-1 | – | – | 97.5%#10 | – | 97.5%#10 | – | ARC Prize | |
| ARC-AGI-2 | – | – | 88.3%#20 | – | 90.4%#12 | – | ARC Prize | |
| ARC-AGI-3 | – | – | 30.2%#13 | – | – | – | ARC Prize | |
| CritPt | 23.1%#48 | 26.9%#34 | 28.3%#25 | 27.7%#27 | 29.1%#19 | – | Artificial Analysis | |
| MysteryMechanismnot in index | – | – | – | – | – | 37.4%#7 | Vals AI | |
| Chess Puzzlesnot in index | 20.0%#85 | – | – | – | 42.0%#21 | 33.0%#45 | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | – | – | – | – | 59.0%#5 | 37.0%#15 | Epoch AI Benchmarking Hub | |
| LMCAnot in index | 62.6%#15 | 62.9%#13 | 63.7%#10 | 64.5%#6 | 63.3%#12 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | 93.1%#52 | 94.9%#34 | 97.9%#3 | 97.9%#3 | 97.6%#5 | – | Epoch AI Benchmarking Hub | |
| BTF-3not in index | – | – | – | 12.0%#13 | – | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SciCode | 48.0%#112 | 50.7%#92 | 54.3%#56 | 55.0%#48 | 56.4%#35 | – | SciCode | |
| WeirdML | – | – | 91.6%#8 | – | 91.8%#7 | 86.3%#13 | WeirdML | |
| FrontierCode | – | – | – | – | 53.4%#3 | – | Cognition | |
| LMArena Coding | – | – | 1533#15 | – | 1530#21 | – | LMArena | |
| LMArena WebDev | – | – | 1660#12 | – | 1695#7 | – | LMArena | |
| LiveBench Codingnot in index | – | – | – | – | 81.5%#14 | – | LiveBench | |
| SciCode (AA)not in index | 49.2%#107 | 51.5%#86 | 55.4%#46 | 55.7%#42 | 56.4%#35 | – | Artificial Analysis | |
| LiveCodeBench | – | – | – | – | – | 89.0%#4 | Vals AI | |
| IOI | – | – | – | – | – | 84.3%#8 | Vals AI | |
| SWE-bench (Vals)not in index | – | – | – | – | – | 97.0%#1 | Vals AI | |
| Code Migrationnot in index | – | – | – | – | – | 57.5%#6 | Vals AI | |
| ProgramBenchnot in index | – | – | – | – | – | 3.0%#4 | Vals AI | |
| Vibe Code Bench 1-100not in index | – | – | – | – | – | 28.5%#2 | Vals AI | |
| Vibe Code Bench v1.1not in index | – | – | – | – | – | 88.4%#8 | Vals AI | |
| SWE Atlas: Codebase QnAnot in index | – | – | – | 63.2%#1 | – | – | Scale AI SEAL | |
| SWE Atlas: Test Writingnot in index | – | – | – | 62.2%#2 | – | – | Scale AI SEAL | |
| DeepSWEnot in index | 58.1%#35 | 68.9%#17 | 72.8%#7 | 73.2%#6 | 73.7%#3 | – | Epoch AI Benchmarking Hub | |
| CursorBenchnot in index | 40.7%#26 | 43.3%#19 | 44.7%#16 | 46.1%#14 | 46.6%#12 | – | Cursor | |
| ALE-Benchnot in index | – | – | 2164.6#4 | – | – | – | Epoch AI Benchmarking Hub | |
| FrontierSWEnot in index | – | – | – | – | 52.0%#6 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | 26.3%#43 | 34.3%#34 | 46.0%#21 | 46.5%#20 | 49.0%#18 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | 76.4%#59 | 86.1%#21 | 87.6%#18 | 88.0%#12 | 89.1%#7 | – | Artificial Analysis |
Agents & tools
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Terminal-Bench | 34.9%#50 | 44.9%#37 | 50.3%#30 | 53.9%#24 | 51.8%#28 | – | Terminal-Bench | |
| OSWorld-Verified 2.0 | 22.3%#6 | 25.3%#5 | 29.0%#3 | 30.2%#2 | 31.4%#1 | – | OSWorld | |
| APEX-Agents | – | – | – | – | 65.8%#6 | – | Mercor | |
| LMArena Agent | – | – | 8.7#7 | – | 7.9#9 | – | LMArena | |
| LiveBench Agentic Codingnot in index | – | – | – | – | 65.2%#5 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | – | – | – | – | – | 84.6%#5 | Vals AI | |
| MCP Atlas | – | – | – | 85.8%#3 | – | – | Scale AI SEAL | |
| HiL-Bench | – | – | – | – | – | 57.0%#2 | Scale AI SEAL | |
| GDPval (AA)not in index | 39.7%#91 | 48.8%#47 | 54.0%#29 | 58.8%#12 | 60.4%#8 | – | Artificial Analysis | |
| τ²-Bench Banking (AA) | 30.3%#63 | 38.6%#36 | 44.7%#16 | 43.3%#19 | 42.1%#23 | – | Artificial Analysis | |
| Analyst Agent (AA)not in index | – | – | – | – | 53.8%#5 | – | Artificial Analysis | |
| CUA-benchnot in index | – | – | – | – | 9.0%#4 | – | Vals AI | |
| CyberBenchnot in index | – | – | – | – | – | 65.4%#26 | Vals AI | |
| Finance Agent v2not in index | – | – | – | – | – | 58.6%#8 | Vals AI | |
| Harvey's Legal Agent Benchmarknot in index | – | – | – | – | – | 6.7%#23 | Vals AI | |
| Legal Research Benchnot in index | – | – | – | – | – | 55.3%#1 | Vals AI | |
| SkillsBenchnot in index | – | – | – | – | – | 60.4%#8 | Vals AI | |
| SREBenchnot in index | – | – | – | – | – | 12.2%#8 | Vals AI | |
| Tax Agent Benchnot in index | – | – | – | – | – | 75.1%#3 | Vals AI | |
| Terminal-Bench 4.0 (Vals)not in index | – | – | – | – | – | 53.5%#7 | Vals AI | |
| Terminal-Bench Sciencenot in index | – | – | – | – | – | 27.1%#7 | Vals AI | |
| Time Horizon Index: KSPnot in index | – | – | – | – | – | 18.8%#4 | Vals AI | |
| BALROGnot in index | – | – | – | – | 63.4%#2 | – | BALROG | |
| PostTrainBenchnot in index | – | – | – | – | – | 35.0%#3 | Epoch AI Benchmarking Hub | |
| GBAEvalnot in index | – | – | – | – | – | 79.6%#1 | Epoch AI Benchmarking Hub | |
| Vending-Bench 2not in index | – | – | – | – | – | 11181.9#4 | Andon Labs | |
| AutomationBenchnot in index | – | 54.3%#32 | 53.6%#34 | 53.2%#35 | 56.6%#29 | – | Artificial Analysis | |
| GDP.pdfnot in index | – | 20.0%#27 | 19.6%#30 | 21.0%#25 | 21.6%#24 | – | Artificial Analysis | |
| Harvey LABnot in index | – | – | – | – | 93.5%#5 | – | Harvey | |
| EnterpriseOps-Gymnot in index | – | – | – | – | 47.5%#7 | – | Artificial Analysis | |
| AA-Briefcasenot in index | – | 1435#32 | 1562#16 | 1636#10 | 1662#7 | – | Artificial Analysis | |
| FACTS Searchnot in index | – | – | – | – | – | 85.3%#5 |
Maths
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | – | – | 85.6%#11 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | – | – | 73.2%#17 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 93.3%#51 | – | – | – | 98.9%#16 | 97.8%#24 | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | – | – | 99.0%#4 | – | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | – | – | 95.7%#14 | – | LiveBench | |
| LMArena Maths | – | – | 1521#3 | – | 1516#7 | – | LMArena |
Knowledge
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | – | 59.9%#16 | – | Epoch AI Benchmarking Hub | |
| LiveBench Data Analysisnot in index | – | – | – | – | 74.5%#35 | – | LiveBench | |
| AA-Omniscience | 28.6#37 | 31.0#29 | 33.7#25 | 35.4#23 | 37.1#22 | – | Artificial Analysis | |
| MMLU-Pro | – | – | – | – | – | 91.6%#2 | Vals AI | |
| LegalBench | – | – | – | – | – | 87.0%#8 | Vals AI | |
| CorpFin | – | – | – | – | – | 73.2%#1 | Vals AI | |
| TaxEval | – | – | – | – | – | 75.1%#22 | Vals AI | |
| BioMysteryBenchnot in index | – | – | – | – | – | 79.3%#3 | Vals AI | |
| Excel Modeling Benchmarknot in index | – | – | – | – | – | 73.6%#6 | Vals AI | |
| MedCodenot in index | – | – | – | – | – | 63.6%#1 | Vals AI | |
| MedScribenot in index | – | – | – | – | – | 91.0%#4 | Vals AI | |
| Public Benefits Benchnot in index | – | – | – | – | – | 76.9%#1 | Vals AI | |
| DrugDiscoveryBenchnot in index | – | – | – | – | 53.7%#4 | – | Scale AI SEAL | |
| EBR-benchnot in index | – | – | – | – | 45.7%#6 | – | Epoch AI Benchmarking Hub | |
| GDP.pdfnot in index | – | – | – | – | 24.0%#15 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 56.0%#34 | 57.1%#31 | 58.9%#24 | 59.5%#22 | 60.9%#15 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 37.8%#160 | 39.3%#148 | 38.8%#153 | 40.5%#144 | 39.2%#149 | – | Artificial Analysis |
Instruction following
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | – | – | 88.7%#5 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | – | – | 63.8%#46 | – | LiveBench | |
| LMArena Instruction Followingnot in index | – | – | 1496#8 | – | 1493#10 | – | LMArena |
Human preference
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Text | – | – | 1490#13 | – | 1490#14 | – | LMArena | |
| EQ-Bench 4 | – | – | – | – | – | 1385#1 | EQ-Bench | |
| LMArena Creative Writingnot in index | – | – | 1471#13 | – | 1471#15 | – | LMArena | |
| LMArena Multi-turnnot in index | – | – | 1484#32 | – | 1487#28 | – | LMArena | |
| LMArena Longer Queriesnot in index | – | – | 1504#10 | – | 1500#14 | – | LMArena |
Multimodal
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| LMArena Vision | – | – | 1321#2 | – | – | – | LMArena | |
| MMMU-Pro | 79.8%#53 | 81.6%#37 | 82.4%#32 | 84.0%#22 | 84.7%#15 | – | Artificial Analysis | |
| MMMU-Pro (Vals)not in index | – | – | – | – | – | 89.9%#2 | Vals AI | |
| MortgageTaxnot in index | – | – | – | – | – | 72.1%#1 | Vals AI | |
| SAGEnot in index | – | – | – | – | – | 49.4%#21 | Vals AI | |
| Blueprint-Bench 2not in index | – | – | – | – | – | 30.4%#16 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| AA-LCR | 81.3%#50 | 82.0%#41 | 79.0%#97 | 80.3%#63 | 79.3%#89 | – | Artificial Analysis | |
| LMArena Documentnot in index | – | – | 1490#7 | – | – | – | LMArena | |
| MLCRnot in index | – | 56.1%#7 | 59.4%#5 | 58.3%#6 | 55.6%#8 | – | Artificial Analysis |
Composite
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in index | – | – | – | – | – | 162.9#5 | Epoch AI Benchmarking Hub | |
| LiveBench | – | – | – | – | 80.1%#12 | – | LiveBench | |
| AA Intelligence Index | 39.4#62 | 44.8#35 | 48.1#22 | 49.7#18 | 50.8#16 | – | Artificial Analysis | |
| Vals Indexnot in index | – | – | – | – | – | 63.7#5 | Vals AI | |
| Vals Multimodal Indexnot in index | – | – | – | – | – | 73.9%#2 | Vals AI |
Safety & honesty
| Benchmark | low | medium | high | xhigh | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| FORTRESSnot in index | – | – | – | – | – | 16.8%#48 | Scale AI SEAL |
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 |
|---|---|---|---|---|---|---|
| Anthropic | 112 tok/s | 1.85 s | $10.00 | $50.00 | 1M | – |
| Google Vertex | 82 tok/s | 2.49 s | $5.50 | $27.50 | 1M | – |
| Google Vertex | 72 tok/s | 2.92 s | $5.50 | $27.50 | 1M | – |
| Google Vertex | 68 tok/s | 2.75 s | $5.00 | $25.00 | 1M | – |
| Claude Platform on AWS | 67 tok/s | 2.40 s | $5.00 | $25.00 | 1M | – |
| Amazon Bedrock | 65 tok/s | 4.60 s | $5.00 | $25.00 | 1M | – |
| Amazon Bedrock | 64 tok/s | 3.76 s | $5.50 | $27.50 | 1M | – |
| Anthropic | 62 tok/s | 3.44 s | $5.00 | $25.00 | 1M | – |
| Amazon Bedrock | 57 tok/s | 4.18 s | $5.50 | $27.50 | 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 | 25.3 s |
| Summarise a 30-page report | 12,000 / 600 | $0.075 | $0.035 | 30.8 s |
| Code edit | 6,000 / 1,500 | $0.068 | $0.047 | 47.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.400 | $0.198 | 1.6 min |
| Structured extraction | 2,000 / 200 | $0.015 | $0.0083 | 23.5 s |
See also
Data as of 5 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Claude Opus 5: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/claude-opus-5, data as of 5 Oct 2026.