Model comparison
Claude 3.7 Sonnet vs Claude Sonnet 4.6
Claude Sonnet 4.6 is the stronger model overall, scoring 50.3 to 39.5 on the Noometry Index.
Last verified . 29 shared benchmarks.
Summary
- They share 29 benchmarks with published results for both. Claude 3.7 Sonnet scores higher in 1 category and Claude Sonnet 4.6 in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Sonnet 4.6 leads 46.1 to 18.6.
- The biggest single-benchmark swing is ARC-AGI-2: 0.9% for Claude 3.7 Sonnet and 60.4% for Claude Sonnet 4.6.
Side by side
| Claude 3.7 Sonnet | Claude Sonnet 4.6 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 39.5 | 50.3 |
| Released | 2025-02-24 | 2026-02-17 |
| Weights | Proprietary | Proprietary |
| Context window | — | 1M |
| Max output | — | 128K |
| Input $ / M tokens | — | $3 |
| Output $ / M tokens | — | $15 |
| Results tracked | 58 | 57 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 40.6 (#136), Claude Sonnet 4.6: 46.3 (#67)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| SWE-bench Verified | 61% | 75.2% |
| LMArena Coding | 1361 | 1504 |
| DeepSWE | — | 29.9% |
| FrontierCode | — | 24.3% |
| SWE-bench Verified (bash only) | 52.8% | — |
| Aider Polyglot | 64.9% | — |
| LMArena WebDev | — | 1522 |
| SciCode | — | 46.8% |
| GSO | 3.8% | — |
| WeirdML | — | 66.1% |
| LiveBench Coding | 74.5% | — |
| CadEval | 54% | — |
| ALE-Bench | — | 1,327 |
Agentic & Tool Use Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 34.1 (#50), Claude Sonnet 4.6: 39.1 (#28)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| DeepResearch Bench | 43.6% | 54.9% |
| OSWorld | 35.8% | 72.1% |
| Terminal-Bench | — | 53.4% |
| APEX-Agents | — | 43% |
| OSWorld 2.0 | — | 9.3% |
| TheAgentCompany | 30.9% | — |
| Cybench | 20% | — |
| ExploitBench | — | 23.6% |
| GBAEval | — | 48.8% |
| GDP.pdf | — | 18% |
| LMArena Search | — | 1221 |
| METR Time Horizons | 60% | — |
| Vending-Bench 2 | — | 7,204 |
Reasoning Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 18.6 (#277), Claude Sonnet 4.6: 46.1 (#45)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| ARC-AGI-2 | 0.9% | 60.4% |
| ARC-AGI-1 | 28.6% | 86.5% |
| LMArena Hard Prompts | 1333 | 1484 |
| Epoch Capabilities Index | 141.16 | 152.24 |
| ForecastBench | 61.8 | 62 |
| SimpleBench | 46.4% | — |
| NYT Connections (extended) | — | 80.9% |
| CritPt | — | 3.1% |
| Chess Puzzles | — | 13% |
| EnigmaEval | 4.2% | — |
| Thematic Generalization | — | 76.3% |
| LiveBench Reasoning | 87.8% | — |
| Mystery Game Puzzles | — | 16% |
| DTBench | — | 89.9% |
| LiveBench Data Analysis | 74% | — |
| LMCA | — | 46.5% |
| LiveBench | 76.1% | — |
Math Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 37.5 (#153), Claude Sonnet 4.6: 52.9 (#49)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 57.8% | 85.8% |
| LMArena Math | 1337 | 1462 |
| FrontierMath (Feb 2025 set) | 4.1% | 32.4% |
| ProofBench | — | 45% |
| Omni-MATH | 33% | — |
| LiveBench Math | 79% | — |
| MATH Level 5 | 91.2% | — |
| FrontierMath Tier 4 (v1) | — | 8.3% |
Knowledge Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 39.8 (#130), Claude Sonnet 4.6: 51.7 (#65)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| GPQA Diamond | 79.7% | 87.4% |
| LMArena Expert | 1321 | 1500 |
| Humanity's Last Exam | 8% | — |
| SimpleQA Verified | — | 35.5% |
| MMLU-Pro | 78.4% | — |
| Confabulations | 14.7% | — |
| Vectara Hallucination Rate | — | 10.6% |
| GPQA (HELM) | 60.8% | — |
Multimodal Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 33.7 (#95), Claude Sonnet 4.6: 38.0 (#68)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Vision | 1169 | 1283 |
| GeoBench | 68% | — |
| VPCT | 39% | — |
| Blueprint-Bench 2 | — | 6.7% |
| LMArena Document | — | 1482 |
| SpatialViz-Bench | 33.9% | — |
Multilingual Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 44.1 (#179), Claude Sonnet 4.6: 54.4 (#41)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Non-English | 1296 | 1440 |
| LMArena Chinese | 1299 | 1491 |
| LMArena French | 1303 | 1465 |
| LMArena German | 1301 | 1428 |
| LMArena Japanese | 1267 | 1420 |
| LMArena Korean | 1249 | 1411 |
| LMArena Russian | 1311 | 1440 |
| LMArena Spanish | 1298 | 1464 |
Instruction Following Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 72.9 (#125), Claude Sonnet 4.6: 77.4 (#25)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Instruction Following | 1352 | 1475 |
| LiveBench Instruction Following | 81.3% | — |
| IFEval | 83.4% | — |
Long Context Claude 3.7 Sonnet leads
Claude 3.7 Sonnet: 50.3 (#10), Claude Sonnet 4.6: 45.3 (#44)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Longer Query | 1373 | 1479 |
| Fiction.LiveBench | 83.3% | — |
Writing & Preference Claude Sonnet 4.6 leads
Claude 3.7 Sonnet: 54.4 (#150), Claude Sonnet 4.6: 70.2 (#22)
| Benchmark | Claude 3.7 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Text | 1314 | 1458 |
| LMArena Creative Writing | 1332 | 1435 |
| EQ-Bench Creative Writing | 1412 | 1810 |
| LMArena Multi-Turn | 1339 | 1464 |
| Short-Story Creative Writing | 81.1% | — |
| WildBench | 81.4% | — |
| EQ-Bench 4 | — | 1207 |
| LiveBench Language | 59.9% | — |
Frequently asked questions
Is Claude 3.7 Sonnet better than Claude Sonnet 4.6?
Claude Sonnet 4.6 is the stronger model overall, scoring 50.3 to 39.5 on the Noometry Index.
Is Claude 3.7 Sonnet or Claude Sonnet 4.6 better for coding?
Claude Sonnet 4.6 scores higher on coding benchmarks: 46.3 versus 40.6 in the Noometry coding category.
How many benchmarks do Claude 3.7 Sonnet and Claude Sonnet 4.6 share?
29 benchmarks have published results for both models. Claude 3.7 Sonnet has 58 scored results on Noometry and Claude Sonnet 4.6 has 57.