Claude Sonnet 4 (unknown thinking): benchmark scores, price and ranking
Claude Sonnet 4 (unknown thinking) from Anthropic ranks #538 of 743 LLMs on the LLMs Tiger aggregate benchmark score (46.1), based on 8 benchmark results captured up to 2026-10-13.
- Organisation
- Anthropic
- Type
- Frontier model
- Open weights
- No
- Released
- 2025-05-22 (source: curated)
- Cheapest API price
- $3.00 in / $15.00 out per million tokens (via github_copilot, observed 2026-10-09T21:47Z)
- Aggregate rank
- #538 of 743 (score 46.1)
Relative to other models, Claude Sonnet 4 (unknown thinking) is strongest on DTBench (#61 of 146) and weakest on CritPt (#76 of 115).
Across the 8 benchmarks where at least ten models are scored, Claude Sonnet 4 (unknown thinking) is in the top 10% on 0 and in the top half on 4.
Its cheapest listed input price, $3.00 per million tokens, is lower than that of 11% of the 427 models with a listed price.
Claude Sonnet 4 (unknown thinking) benchmark results
| Benchmark | Score | Rank | Source | Captured |
|---|---|---|---|---|
| DTBench | 77.1% | #61 of 146 | Epoch AI Benchmarking Hub | 2026-10-13 |
| GSO-Bench | 4.9% | #8 of 19 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SimpleBench | 45.5% | #27 of 61 | Epoch AI Benchmarking Hub | 2026-10-13 |
| EnigmaEval | 3.1% | #12 of 24 | Epoch AI Benchmarking Hub | 2026-10-13 |
| LMCA | 29.0% | #56 of 108 | Epoch AI Benchmarking Hub | 2026-10-13 |
| SciCode | 40.0% | #65 of 106 | Epoch AI Benchmarking Hub | 2026-10-13 |
| Humanity's Last Exam | 7.8% | #17 of 27 | Epoch AI Benchmarking Hub | 2026-10-13 |
| CritPt | 0.3% | #76 of 115 | Epoch AI Benchmarking Hub | 2026-10-13 |
About Claude Sonnet 4 (unknown thinking)
Claude Sonnet 4 (unknown thinking) is a frontier model from Anthropic. It was released on 2025-05-22, according to the developer's announcement. The cheapest API listing we track charges $3.00 per million input tokens and $15.00 per million output tokens, through github_copilot. We hold 8 benchmark results for it across 4 areas, and none is in the top tenth of its board.
Where Claude Sonnet 4 (unknown thinking) stands
On our aggregate score, Claude Sonnet 4 (unknown thinking) is #538 of 743 ranked models, in the lower half. Its score is 46.1; the scale is centred on 50, so it sits 3.9 points below the typical model. That beats 28% of the ranked field. Among the 70 ranked Anthropic models it is #57. The next one up is Claude Sonnet 3.5 at 46.7. The next one down is claude-opus-4-1-20250805_32K at 46.0.
Claude Sonnet 4 (unknown thinking), benchmark by benchmark
DTBench puts Claude Sonnet 4 (unknown thinking) at 77.1%, #61 of 146. That beats 58% of the models tested, but 21.9 points short of Claude Opus 5.5 (max) (98.9%). The results either side are o4-mini (high) (77.6%) and Qwen 3.6 Flash (2026-04-16) (77.1%). It covers dynamic tool-use benchmark. The board median is 68.8%, so Claude Sonnet 4 (unknown thinking) is 8.3 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
On GSO-Bench, Claude Sonnet 4 (unknown thinking) reaches 4.9%. It ranks #8 of 19, ahead of 58% of the field, and 35.3 points behind GPT-5.5 (xhigh) at 40.2%. The results either side are GPT-5 (high) (6.9%) and claude-sonnet-4-20250514 (4.9%). It covers software performance optimisation. The median score on this board is also 4.9%. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Claude Sonnet 4 (unknown thinking) manages 45.5% on SimpleBench, good for #27 of 61. Claude Fable 5 (max) leads at 81.9%, 36.4 points further up, and 56% of models sit below. The results either side are MiniMax-M3 (45.8%) and claude-sonnet-4-20250514_12K (45.5%). It covers trick questions that test common sense. The board median is 41.4%, so Claude Sonnet 4 (unknown thinking) is 4.1 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
EnigmaEval puts Claude Sonnet 4 (unknown thinking) at 3.1%, #12 of 24. That beats 50% of the models tested, but 15.1 points short of Gemini 3 Pro Preview (18.2%). The results either side are GPT-4.5 Preview (Feb 2025) (3.2%) and gemini-2.5-flash-preview-05-20 (2.7%). It covers multi-step puzzle hunts. The board median is 2.9%, so Claude Sonnet 4 (unknown thinking) is 0.2 points above it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Claude Sonnet 4 (unknown thinking) lands below the middle on LMCA: 29.0%, #56 of 108, with 51% of the field scoring higher. Claude Opus 5.5 (max) leads at 68.2%; the closest result above is Qwen3.5 Flash at 29.1%. It covers language-model capability assessment. The board median is 29.2%, so Claude Sonnet 4 (unknown thinking) is 0.2 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Claude Sonnet 4 (unknown thinking) lands below the middle on SciCode: 40.0%, #65 of 106, with 60% of the field scoring higher. Claude Opus 5.5 (max) leads at 66.9%; the closest result above is Mistral Medium 3.5 at 40.2%. It covers scientific research code. The board median is 44.3%, so Claude Sonnet 4 (unknown thinking) is 4.3 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Claude Sonnet 4 (unknown thinking) lands below the middle on Humanity's Last Exam: 7.8%, #17 of 27, with 59% of the field scoring higher. Muse Spark leads at 40.6%; the closest result above is GLM-4.5-Air at 8.1%. It covers expert-level questions across domains. The board median is 10.7%, so Claude Sonnet 4 (unknown thinking) is 3.0 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Claude Sonnet 4 (unknown thinking) lands below the middle on CritPt: 0.3%, #76 of 115, with 65% of the field scoring higher. GPT-5.6 Sol (max) leads at 32.3%; the closest result above is Claude Opus 4 at 0.3%. It covers unpublished research-level physics problems. The board median is 2.9%, so Claude Sonnet 4 (unknown thinking) is 2.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.
Strengths and weaknesses
Software engineering: uneven, ahead of 48% of the field on average across 2 benchmarks. Its best board here is GSO-Bench (#8 of 19) and its weakest is SciCode (#65 of 106). Reasoning, maths and science: uneven, ahead of 44% of the field on average across 4 benchmarks. Its best board here is SimpleBench (#27 of 61) and its weakest is CritPt (#76 of 115).
Price and value
The median input price across the 427 priced models is $0.500 per million tokens, so Claude Sonnet 4 (unknown thinking) costs more than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on gpt-5.5-xhigh (codex-harness) ($5.00) and GPT-4 Turbo Preview (Nov 2023) ($10.00).
Release history
The Anthropic model released before it was claude-sonnet-4-20250514_32K on 2025-05-22, which scores 49.9. The next one was claude-opus-4-1-20250805 on 2025-08-05, at 49.7.
Reading these numbers
The 8 scores on this page come from 1 source: Epoch AI Benchmarking Hub (8). No official leaderboard supplied any of them, so treat them as third-party measurements. The leaderboard also lists 4 other reasoning-effort settings of this model, which redirect here. This page shows the highest.
Claude Sonnet 4 (unknown thinking) is also listed at 4 other reasoning-effort settings; this page shows the highest-effort row, which is the one the leaderboard keeps by default.
Models ranked nearby
- #535 Qwen3.6 35B-A3B (46.2)
- #536 Phi-3-mini-4k-instruct (46.2)
- #537 Gemini 2.0 Flash (Feb 2025) (46.1)
- #539 gemma2:27b-instruct-q8_0 (46.1)
- #540 Mathstral-7B-v0.1 (46.1)
- #541 MiMo-7B-Base (46.0)
More from Anthropic
- Claude Opus 5.5 (max) (#1, 64.3)
- Claude Fable 5.1 (max) (#3, 61.9)
- Claude Sonnet 5.5 (max) (#4, 60.8)
- Claude Opus 5 (max) (#8, 60.0)
- Claude Fable 5 (max) (#9, 59.9)
- claude-opus-4-8-xhigh-effort (#14, 56.6)
Compare it on the live leaderboard → · How the aggregate score works