AnthropicClaude Sonnet 4.6
Reasoning effort
Claude Sonnet 4.6 is an Anthropic proprietary model, released 17 Feb 2026. Its best configuration ranks #185 of 785 on the BenchLeader Index at 57.1 ±3.3, in the upper half. The ± is the point: 250 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 reasoning (64) and lowest in composite (47). At $6.00 per million tokens blended it is among the most expensive ranked models. Output speed of 44 tokens per second puts it in the slowest quarter, with a first token in 0.7 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 22 Oct 2026.
- Blended price
- $6.00/M
- $3.00 in · $15.00 out
- Output speed
- 44 tok/s
- OpenRouter traffic, 7-day median; not yet measured by Artificial Analysis
- First answer
- 0.74 s
- Context
- 1M
- Full answer
- –
- Cost per run
- –
- Released
- 17 Feb 2026
- Overall index57
- Reasoning64
- Coding56
- Agents & tools51
- Maths63
- Knowledge60
- Human preference61
- Multimodal62
- Composite47
Versions
Anthropic has shipped 10 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 Sonnet 5.5max | 28 Sept 2026 | 69.3 | #13 |
| Claude Sonnet 5high | 30 Jun 2026 | 60.2 | #114 |
| Claude Sonnet 4.6this page | 17 Feb 2026 | 57.1 | #185 |
| Claude Sonnet 4.5high | 29 Sept 2025 | 58.1 | #160 |
| Claude Sonnet 4thinking | 22 May 2025 | 53.2 | #294 |
| Claude 3.7 Sonnetthinking | 24 Feb 2025 | 51.0 | #353 |
| Claude 3.5 Sonnet | 22 Oct 2024 | 46.3 | #490 |
| Claude 3 Sonnet | 29 Feb 2024 | 38.0 | #721 |
| Claude 37 Sonnetthinking | – | – | – |
| Claude Sonnet 5.0max | – | – | – |
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 |
|---|---|---|---|---|---|---|
| no reasoning | 54.0 | #263 | 47 tok/s | 0.98 s | $0.0057 | Agents & tools 74 · Composite 57 · Instruction following 44 · Knowledge 60 · Long context 59 · Multimodal 53 · Reasoning 42 |
| low | 53.5 | #286 | 46 tok/s | 1.07 s | $0.0057 | Agents & tools 71 · Composite 56 · Instruction following 45 · Knowledge 61 · Long context 60 · Multimodal 52 · Reasoning 42 |
| medium | – | – | 44 tok/s | 0.74 s | $0.0057 | Agents & tools 38 · Coding 61 · Maths 59 · Reasoning 60 |
| high | 50.7 | #364 | 44 tok/s | 0.74 s | $0.0057 | Agents & tools 43 · Knowledge 45 · Maths 56 · Reasoning 60 |
| max | 56.0 | #215 | 56 tok/s | 64 s | $0.0057 | Agents & tools 61 · Coding 50 · Composite 63 · Instruction following 57 · Knowledge 55 · Long context 65 · Maths 53 · Multimodal 56 · Reasoning 55 |
| not statedbest | 57.1 | #185 | 44 tok/s | 0.74 s | $0.0057 | Agents & tools 51 · Coding 56 · Composite 47 · Human preference 61 · Knowledge 60 · Maths 63 · Multimodal 62 · Reasoning 64 |
How its index has moved
8 Sept 2026 to 23 Oct 2026- not stated
- max
- no reasoning
- low
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
See all LLM benchmarks. 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 | low | medium | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| GPQA Diamond | – | – | 83.3%#103 | 83.3%#103 | 78.8%#132 | 87.4%#73 | Epoch AI Benchmarking Hub | |
| Arena Hard Promptseffort not stated | – | – | – | – | – | 1503#28 | Arena (LMArena) | |
| LiveBench 2.0 Reasoningnot in indexeffort not stated | – | – | – | – | – | 86.4%#28 | LiveBench | |
| LiveBench 1.0 Reasoningnot in index | – | – | 84.8%#39 | – | – | – | LiveBench | |
| GPQA Diamond (AA)not in index | 79.9%#195 | 79.7%#196 | – | – | 87.5%#97 | – | Artificial Analysis | |
| Humanity's Last Exam (AA)not in index | 13.3%#271 | 11.2%#297 | – | – | 33.6%#135 | – | Artificial Analysis | |
| GPQA Diamond (Vals)not in indexeffort not stated | – | – | – | – | – | 85.6%#44 | Vals AI | |
| ARC-AGI-1 | – | – | – | 86.5%#104 | 86.0%#107 | – | ARC Prize | |
| ARC-AGI-2 | – | – | – | 60.4%#86 | 58.3%#94 | – | ARC Prize | |
| CritPt | 0.9%#237 | 0.9%#237 | – | – | 3.1%#176 | – | Artificial Analysis | |
| Chess Puzzlesnot in index | – | – | 16.0%#104 | 19.0%#95 | 23.0%#82 | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | – | 16.0%#84 | – | – | – | 14.0%#97 | Epoch AI Benchmarking Hub | |
| LMCAnot in index | – | – | – | – | 46.5%#83 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | – | 89.9%#88 | – | Epoch AI Benchmarking Hub | |
| ForecastBenchnot in indexeffort not stated | – | – | – | – | – | 64.2%#31 | ForecastBench |
Coding
| Benchmark | medium | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| SWE-bench Verified (Epoch)effort not stated | – | – | – | 75.2%#14 | Epoch AI Benchmarking Hub | |
| SciCode | – | – | 46.8%#130 | – | SciCode | |
| WeirdML | 66.1%#40 | – | – | – | WeirdML | |
| FrontierCodeeffort not stated | – | – | – | 24.3%#38 | Cognition | |
| Arena Codingeffort not stated | – | – | – | 1528#24 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | – | – | 1523#51 | Arena (LMArena) | |
| LiveBench 2.0 Codingnot in indexeffort not stated | – | – | – | 66.6%#42 | LiveBench | |
| LiveBench 1.0 Codingnot in index | 79.3%#25 | – | – | – | LiveBench | |
| SciCode (AA)not in index | – | – | 50.1%#110 | – | Artificial Analysis | |
| LiveCodeBencheffort not stated | – | – | – | 82.1%#55 | Vals AI | |
| SWE-bench (Vals)not in indexeffort not stated | – | – | – | 77.4%#34 | Vals AI | |
| Code Migrationnot in indexeffort not stated | – | – | – | 39.9%#30 | Vals AI | |
| ProgramBenchnot in indexeffort not stated | – | – | – | 0.5%#18 | Vals AI | |
| Vibe Code Bench v1.1not in indexeffort not stated | – | – | – | 51.5%#59 | Vals AI | |
| SWE Atlas: Codebase QnAnot in indexeffort not stated | – | – | – | 31.2%#15 | Scale AI SEAL | |
| SWE Atlas: Refactoringnot in indexeffort not stated | – | – | – | 32.2%#13 | Scale AI SEAL | |
| SWE Atlas: Test Writingnot in indexeffort not stated | – | – | – | 31.8%#15 | Scale AI SEAL | |
| DeepSWE v1.1not in index | – | 29.9%#64 | – | – | Epoch AI Benchmarking Hub | |
| ALE-Benchnot in index | 1327.3#29 | – | – | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in index | – | – | 3.0%#108 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in index | – | – | 71.2%#71 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | low | medium | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | – | – | – | – | 53.4%#29 | Terminal-Bench | |
| OSWorld-Verified 2.0 | – | – | 9.3%#10 | – | 8.3%#11 | – | OSWorld | |
| APEX-Agents | – | – | – | 43.0%#40 | – | – | Mercor | |
| LiveBench Long Agenticnot in indexeffort not stated | – | – | – | – | – | 23.1%#40 | LiveBench | |
| LiveBench Browsernot in indexeffort not stated | – | – | – | – | – | 68.0%#24 | LiveBench | |
| LiveBench Agentic Codingnot in index | – | – | 42.6%#58 | – | – | – | LiveBench | |
| Terminal-Bench Hard | 46.2%#30 | 42.4%#46 | – | – | 53.0%#17 | – | Artificial Analysis | |
| τ²-Bench Telecom (AA)not in index | 79.5%#126 | 79.0%#129 | – | – | 75.7%#136 | – | Artificial Analysis | |
| Terminal-Bench 2.1 (Vals)effort not stated | – | – | – | – | – | 57.3%#43 | Vals AI | |
| MCP Atlaseffort not stated | – | – | – | – | – | 69.5%#22 | Scale AI SEAL | |
| τ³-Banking (AA) | – | – | – | – | 34.4%#52 | – | Artificial Analysis | |
| ITBench SRE (AA)not in index | – | – | – | – | 39.8%#21 | – | Artificial Analysis | |
| Analyst Agent (AA)not in index | – | – | – | – | 20.0%#34 | – | Artificial Analysis | |
| APEX-Agents (AA)not in index | – | – | – | – | 28.0%#15 | – | Artificial Analysis | |
| Finance Agent v2not in indexeffort not stated | – | – | – | – | – | 51.0%#41 | Vals AI | |
| Harvey's Legal Agent Benchmark v1not in indexeffort not stated | – | – | – | – | – | 5.0%#37 | Vals AI | |
| Legal Research Benchnot in indexeffort not stated | – | – | – | – | – | 38.5%#34 | Vals AI | |
| SkillsBenchnot in indexeffort not stated | – | – | – | – | – | 49.0%#23 | Vals AI | |
| Tax Agent Benchnot in indexeffort not stated | – | – | – | – | – | 25.8%#37 | Vals AI | |
| Terminal-Bench 2.0 (Vals)not in indexeffort not stated | – | – | – | – | – | 59.5%#7 | Vals AI | |
| DeepResearch Benchnot in index | – | 50.4%#9 | – | 54.9%#2 | – | – | DeepResearch Bench | |
| ExploitBenchnot in indexeffort not stated | – | – | – | – | – | 23.6%#5 | Epoch AI Benchmarking Hub | |
| GBAEvalnot in indexeffort not stated | – | – | – | – | – | 48.8%#8 | Epoch AI Benchmarking Hub | |
| Vending-Bench 2not in indexeffort not stated | – | – | – | – | – | 7204.1#17 | Andon Labs | |
| AutomationBenchnot in index | – | – | – | – | 20.1%#137 | – | Artificial Analysis | |
| GDP.pdf (AA)not in index | – | – | – | – | 15.8%#92 | – | Artificial Analysis |
Maths
| Benchmark | medium | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| OTIS Mock AIME | 82.2%#108 | 75.6%#126 | 71.1%#135 | 85.8%#95 | Epoch AI Benchmarking Hub | |
| ProofBench | – | – | 45.0%#35 | – | Vals AI | |
| LiveBench Mathematicsnot in index | 87.0%#48 | – | – | – | LiveBench | |
| AIME (Vals)effort not stated | – | – | – | 92.3%#20 | Vals AI | |
| Arena Mathseffort not stated | – | – | – | 1463#62 | Arena (LMArena) |
Knowledge
| Benchmark | no reasoning | low | medium | high | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|---|
| SimpleQA Verified | – | – | – | 35.5%#51 | 32.8%#61 | – | Epoch AI Benchmarking Hub | |
| LiveBench 2.0 Data Analysisnot in indexeffort not stated | – | – | – | – | – | 76.1%#27 | LiveBench | |
| LiveBench Researchnot in indexeffort not stated | – | – | – | – | – | 63.6%#41 | LiveBench | |
| LiveBench 1.0 Data Analysisnot in index | – | – | 78.0%#29 | – | – | – | LiveBench | |
| AA-Omniscience | -3.5#153 | -2.1#147 | – | – | 12.2#90 | – | Artificial Analysis | |
| MMLU-Proeffort not stated | – | – | – | – | – | 87.3%#29 | Vals AI | |
| LegalBencheffort not stated | – | – | – | – | – | 82.1%#73 | Vals AI | |
| CorpFineffort not stated | – | – | – | – | – | 65.3%#29 | Vals AI | |
| TaxEvaleffort not stated | – | – | – | – | – | 77.1%#4 | Vals AI | |
| MedQAeffort not stated | – | – | – | – | – | 92.1%#36 | Vals AI | |
| CaseLaw v2not in indexeffort not stated | – | – | – | – | – | 64.0%#14 | Vals AI | |
| Excel Modeling Benchmarknot in indexeffort not stated | – | – | – | – | – | 60.1%#35 | Vals AI | |
| Public Benefits Bench v1.1not in indexeffort not stated | – | – | – | – | – | 62.5%#26 | Vals AI | |
| DrugDiscoveryBenchnot in index | – | – | – | – | 31.3%#15 | – | Scale AI SEAL | |
| GDP.pdfnot in index | – | – | – | – | 18.0%#33 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in index | 38.5%#135 | 36.5%#148 | – | – | 40.9%#118 | – | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in index | 31.5%#213 | 39.2%#164 | – | – | 51.5%#111 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | low | medium | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| LiveBench Languagenot in index | – | – | 76.1%#46 | – | – | LiveBench | |
| LiveBench Instruction Followingnot in index | – | – | 63.2%#51 | – | – | LiveBench | |
| IFBench | 41.2%#253 | 42.4%#238 | – | 56.6%#142 | – | Artificial Analysis | |
| Arena Instruction Followingnot in indexeffort not stated | – | – | – | – | 1475#28 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1472#45 | Arena (LMArena) | |
| EQ-Bench 4 | 1207#15 | EQ-Bench | |
| Arena Creative Writingnot in index | 1449#40 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1481#38 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1492#24 | Arena (LMArena) |
Multimodal
| Benchmark | no reasoning | low | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| Arena Visioneffort not stated | – | – | – | 1283#33 | Arena (LMArena) | |
| LiveBench Designnot in indexeffort not stated | – | – | – | 71.8%#36 | LiveBench | |
| MMMU-Pro | 70.6%#141 | 69.2%#156 | 73.3%#124 | – | Artificial Analysis | |
| MMMU-Pro (Vals)not in indexeffort not stated | – | – | – | 83.6%#30 | Vals AI | |
| MortgageTaxnot in indexeffort not stated | – | – | – | 67.7%#22 | Vals AI | |
| SAGEnot in indexeffort not stated | – | – | – | 46.6%#34 | Vals AI | |
| Blueprint-Bench 2not in indexeffort not stated | – | – | – | 6.7%#27 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | no reasoning | low | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| AA-LCR | 68.3%#235 | 69.3%#226 | 80.0%#78 | – | Artificial Analysis | |
| Arena Documentnot in indexeffort not stated | – | – | – | 1478#13 | Arena (LMArena) | |
| MLCRnot in index | – | – | 7.8%#128 | – | Artificial Analysis |
Composite
| Benchmark | no reasoning | low | medium | max | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | – | – | – | 152.8#48 | Epoch AI Benchmarking Hub | |
| LiveBench 2.0effort not stated | – | – | – | – | 65.1%#32 | LiveBench | |
| LiveBench 1.0not in index | – | – | 73.0%#48 | – | – | LiveBench | |
| AA Intelligence Index v4.3.2 | 24.7#172 | 23.3#188 | – | 30.1#125 | – | Artificial Analysis | |
| Vals Multimodal Indexnot in indexeffort not stated | – | – | – | – | 60.6%#17 | 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 |
|---|---|---|---|---|---|---|
| Amazon Bedrock | 87 tok/s | 0.88 s | $3.30 | $16.50 | 1M | – |
| Amazon Bedrock | 60 tok/s | 0.92 s | $3.00 | $15.00 | 1M | – |
| Claude Platform on AWS | 44 tok/s | 0.68 s | $3.00 | $15.00 | 1M | – |
| Anthropic | 35 tok/s | 1.24 s | $3.00 | $15.00 | 1M | – |
| Google Vertex | 29 tok/s | 1.20 s | $3.00 | $15.00 | 1M | – |
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 $0.300 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0057 | $0.0049 | 7.6 s |
| Summarize a 30-page report | 12,000 / 600 | $0.045 | $0.021 | 14.4 s |
| Code edit | 6,000 / 1,500 | $0.041 | $0.028 | 34.8 s |
| Agentic coding session | 60,000 / 4,000 | $0.240 | $0.118 | 1.5 min |
| Structured extraction | 2,000 / 200 | $0.0090 | $0.0050 | 5.3 s |
More from Anthropic
See also
Questions people ask
- What are Claude Sonnet 4.6's benchmark scores?
- As of 24 Oct 2026, Claude Sonnet 4.6 scores 57.1 ±3.3 on the BenchLeader Index, strongest in reasoning (64) and weakest in composite (47). That rests on 22 independent benchmark results, and every score on this page links to its publisher.
- How does Claude Sonnet 4.6 rank?
- Claude Sonnet 4.6 is #185 of 785 ranked configurations on 24 Oct 2026. 250 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
- How much does Claude Sonnet 4.6 cost per 1M tokens?
- As of 24 Oct 2026, Claude Sonnet 4.6 costs $3.00 per 1M input tokens and $15.00 per 1M output tokens, $6.00 blended (three input tokens per output token).
- When was Claude Sonnet 4.6 released?
- Anthropic released Claude Sonnet 4.6 on 17 Feb 2026. BenchLeader has results for it at 6 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Claude Sonnet 4.6: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/claude-sonnet-4-6.