Model comparison
Claude 3 Sonnet vs Claude Sonnet 4.6
Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 28.8 on the Noometry Index.
Last verified . 23 shared benchmarks.
Summary
- They share 23 benchmarks with published results for both. Claude 3 Sonnet scores higher in 0 categories and Claude Sonnet 4.6 in 9 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 4.6 leads 51.9 to 10.0.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 2.5% for Claude 3 Sonnet and 85.8% for Claude Sonnet 4.6.
Side by side
| Claude 3 Sonnet | Claude Sonnet 4.6 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 28.8 | 50.1 |
| Released | 2024-02-29 | 2026-02-17 |
| Weights | Proprietary | Proprietary |
| Context window | — | 1M |
| Max output | — | 128K |
| Input $ / M tokens | — | $3 |
| Output $ / M tokens | — | $15 |
| Results tracked | 30 | 57 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4.6 leads
Claude 3 Sonnet: 29.2 (#296), Claude Sonnet 4.6: 46.0 (#62)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| WeirdML | 10.2% | 66.1% |
| LMArena Coding | 1223 | 1504 |
| SWE-bench Verified | — | 75.2% |
| DeepSWE | — | 29.9% |
| FrontierCode | — | 24.3% |
| LMArena WebDev | — | 1521 |
| SciCode | — | 46.8% |
| BigCodeBench Instruct | 42.7% | — |
| BigCodeBench Complete | 53.8% | — |
| ALE-Bench | — | 1,327 |
| HumanEval+ | 64% | — |
| MBPP+ | 69.3% | — |
Agentic & Tool Use Not comparable
Claude 3 Sonnet: —, Claude Sonnet 4.6: 39.1 (#26)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| Terminal-Bench | — | 53.4% |
| APEX-Agents | — | 43% |
| OSWorld 2.0 | — | 9.3% |
| DeepResearch Bench | — | 54.9% |
| OSWorld | — | 72.1% |
| ExploitBench | — | 23.6% |
| GBAEval | — | 48.8% |
| GDP.pdf | — | 18% |
| LMArena Search | — | 1221 |
| Vending-Bench 2 | — | 7,204 |
Reasoning Claude Sonnet 4.6 leads
Claude 3 Sonnet: 19.7 (#232), Claude Sonnet 4.6: 46.1 (#40)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Hard Prompts | 1197 | 1485 |
| DTBench | 53.6% | 89.9% |
| Epoch Capabilities Index | 120.69 | 152.25 |
| ARC-AGI-2 | — | 60.4% |
| NYT Connections (extended) | — | 80.9% |
| ARC-AGI-1 | — | 86.5% |
| CritPt | — | 3.1% |
| Chess Puzzles | — | 13% |
| Thematic Generalization | — | 76.3% |
| Mystery Game Puzzles | — | 16% |
| LMCA | — | 46.5% |
| ForecastBench | — | 62 |
| WinoGrande | 75.1% | — |
Math Claude Sonnet 4.6 leads
Claude 3 Sonnet: 10.0 (#305), Claude Sonnet 4.6: 51.9 (#41)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 2.5% | 85.8% |
| LMArena Math | 1213 | 1463 |
| ProofBench | — | 45% |
| MATH Level 5 | 18.2% | — |
| FrontierMath (Feb 2025 set) | — | 32.4% |
| FrontierMath Tier 4 (v1) | — | 8.3% |
Knowledge Claude Sonnet 4.6 leads
Claude 3 Sonnet: 21.1 (#271), Claude Sonnet 4.6: 51.6 (#60)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| GPQA Diamond | 40.6% | 87.4% |
| LMArena Expert | 1173 | 1500 |
| SimpleQA Verified | — | 35.5% |
| Vectara Hallucination Rate | — | 10.6% |
| MMLU | 75.9% | — |
Multimodal Claude Sonnet 4.6 leads
Claude 3 Sonnet: 24.7 (#120), Claude Sonnet 4.6: 38.0 (#61)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Vision | 984 | 1283 |
| Blueprint-Bench 2 | — | 6.7% |
| LMArena Document | — | 1482 |
Multilingual Claude Sonnet 4.6 leads
Claude 3 Sonnet: 38.0 (#229), Claude Sonnet 4.6: 54.6 (#39)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Non-English | 1205 | 1440 |
| LMArena Chinese | 1188 | 1491 |
| LMArena French | 1228 | 1465 |
| LMArena German | 1204 | 1428 |
| LMArena Japanese | 1131 | 1419 |
| LMArena Korean | 1128 | 1411 |
| LMArena Russian | 1227 | 1440 |
| LMArena Spanish | 1204 | 1465 |
Instruction Following Claude Sonnet 4.6 leads
Claude 3 Sonnet: 63.1 (#230), Claude Sonnet 4.6: 77.6 (#23)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Instruction Following | 1200 | 1476 |
Long Context Claude Sonnet 4.6 leads
Claude 3 Sonnet: 36.7 (#223), Claude Sonnet 4.6: 45.3 (#40)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Longer Query | 1211 | 1479 |
Writing & Preference Claude Sonnet 4.6 leads
Claude 3 Sonnet: 42.3 (#233), Claude Sonnet 4.6: 70.4 (#22)
| Benchmark | Claude 3 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| LMArena Text | 1218 | 1458 |
| LMArena Creative Writing | 1186 | 1435 |
| LMArena Multi-Turn | 1227 | 1464 |
| EQ-Bench Creative Writing | — | 1810 |
| EQ-Bench 4 | — | 1207 |
Frequently asked questions
Is Claude 3 Sonnet better than Claude Sonnet 4.6?
Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 28.8 on the Noometry Index.
Is Claude 3 Sonnet or Claude Sonnet 4.6 better for coding?
Claude Sonnet 4.6 scores higher on coding benchmarks: 46.0 versus 29.2 in the Noometry coding category.
How many benchmarks do Claude 3 Sonnet and Claude Sonnet 4.6 share?
23 benchmarks have published results for both models. Claude 3 Sonnet has 30 scored results on Noometry and Claude Sonnet 4.6 has 57.