Model comparison
Claude Sonnet 4.6 vs Claude Sonnet 5
Claude Sonnet 5 is the stronger model overall, scoring 54.6 to 50.3 on the Noometry Index.
Last verified . 44 shared benchmarks.
Summary
- They share 44 benchmarks with published results for both. Claude Sonnet 4.6 scores higher in 4 categories and Claude Sonnet 5 in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Sonnet 5 leads 66.2 to 52.9.
- The biggest single-benchmark swing is ProofBench: 45% for Claude Sonnet 4.6 and 77% for Claude Sonnet 5.
- Claude Sonnet 5 is cheaper at $2 / $10 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.6.
Side by side
| Claude Sonnet 4.6 | Claude Sonnet 5 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 50.3 | 54.6 |
| Released | 2026-02-17 | 2026-06-29 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1M |
| Max output | 128K | 128K |
| Input $ / M tokens | $3 | $2 |
| Output $ / M tokens | $15 | $10 |
| Results tracked | 57 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 5 leads
Claude Sonnet 4.6: 46.3 (#67), Claude Sonnet 5: 55.5 (#26)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| DeepSWE | 29.9% | 53.8% |
| FrontierCode | 24.3% | 42.7% |
| LMArena WebDev | 1522 | 1541 |
| SciCode | 46.8% | 54.3% |
| WeirdML | 66.1% | 68.8% |
| LMArena Coding | 1504 | 1483 |
| ALE-Bench | 1,327 | 1,463 |
| SWE-bench Verified | 75.2% | — |
| CursorBench | — | 34.1% |
| GSO | — | 37.3% |
Agentic & Tool Use Claude Sonnet 5 leads
Claude Sonnet 4.6: 39.1 (#28), Claude Sonnet 5: 42.8 (#18)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| APEX-Agents | 43% | 54.5% |
| GBAEval | 48.8% | 65.3% |
| LMArena Search | 1221 | 1194 |
| Vending-Bench 2 | 7,204 | 6,378 |
| Terminal-Bench | 53.4% | — |
| OSWorld 2.0 | 9.3% | — |
| DeepResearch Bench | 54.9% | — |
| OSWorld | 72.1% | — |
| ExploitBench | 23.6% | — |
| GDP.pdf | 18% | — |
Reasoning Claude Sonnet 5 leads
Claude Sonnet 4.6: 46.1 (#45), Claude Sonnet 5: 49.1 (#39)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| NYT Connections (extended) | 80.9% | 75.1% |
| CritPt | 3.1% | 16.9% |
| Chess Puzzles | 13% | 35% |
| LMArena Hard Prompts | 1484 | 1461 |
| Mystery Game Puzzles | 16% | 35% |
| DTBench | 89.9% | 92.5% |
| LMCA | 46.5% | 50% |
| Epoch Capabilities Index | 152.24 | 156.21 |
| ForecastBench | 62 | 61.1 |
| ARC-AGI-2 | 60.4% | — |
| SimpleBench | — | 60.6% |
| ARC-AGI-1 | 86.5% | — |
| Thematic Generalization | 76.3% | — |
| Surface Evolver Bench | — | 60% |
| Bench to the Future 3 | — | 0.14 |
Math Claude Sonnet 5 leads
Claude Sonnet 4.6: 52.9 (#49), Claude Sonnet 5: 66.2 (#27)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 85.8% | 94.7% |
| ProofBench | 45% | 77% |
| LMArena Math | 1462 | 1467 |
| FrontierMath (Tiers 1-3) | — | 65.6% |
| FrontierMath Tier 4 | — | 29.3% |
| FrontierMath (Feb 2025 set) | 32.4% | — |
| FrontierMath Tier 4 (v1) | 8.3% | — |
Knowledge Claude Sonnet 5 leads
Claude Sonnet 4.6: 51.7 (#65), Claude Sonnet 5: 55.6 (#47)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| GPQA Diamond | 87.4% | 90.5% |
| SimpleQA Verified | 35.5% | 33.7% |
| LMArena Expert | 1500 | 1490 |
| Vectara Hallucination Rate | 10.6% | — |
Multimodal Claude Sonnet 5 leads
Claude Sonnet 4.6: 38.0 (#68), Claude Sonnet 5: 42.4 (#31)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| LMArena Vision | 1283 | 1274 |
| Blueprint-Bench 2 | 6.7% | 24.9% |
| LMArena Document | 1482 | 1466 |
Multilingual Too close to call
Claude Sonnet 4.6: 54.4 (#41), Claude Sonnet 5: 53.8 (#55)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| LMArena Non-English | 1440 | 1431 |
| LMArena Chinese | 1491 | 1477 |
| LMArena French | 1465 | 1460 |
| LMArena German | 1428 | 1440 |
| LMArena Japanese | 1420 | 1422 |
| LMArena Korean | 1411 | 1411 |
| LMArena Russian | 1440 | 1451 |
| LMArena Spanish | 1464 | 1437 |
Instruction Following Claude Sonnet 4.6 leads
Claude Sonnet 4.6: 77.4 (#25), Claude Sonnet 5: 76.3 (#41)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| LMArena Instruction Following | 1475 | 1452 |
Long Context Too close to call
Claude Sonnet 4.6: 45.3 (#44), Claude Sonnet 5: 44.8 (#55)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| LMArena Longer Query | 1479 | 1463 |
Writing & Preference Too close to call
Claude Sonnet 4.6: 70.2 (#22), Claude Sonnet 5: 69.2 (#25)
| Benchmark | Claude Sonnet 4.6 | Claude Sonnet 5 |
|---|---|---|
| LMArena Text | 1458 | 1442 |
| LMArena Creative Writing | 1435 | 1416 |
| EQ-Bench Creative Writing | 1810 | 1794 |
| EQ-Bench 4 | 1207 | 1236 |
| LMArena Multi-Turn | 1464 | 1454 |
Frequently asked questions
Is Claude Sonnet 4.6 better than Claude Sonnet 5?
Claude Sonnet 5 is the stronger model overall, scoring 54.6 to 50.3 on the Noometry Index.
Which is cheaper, Claude Sonnet 4.6 or Claude Sonnet 5?
Claude Sonnet 5 is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; Claude Sonnet 4.6 lists at $3 and $15.
Is Claude Sonnet 4.6 or Claude Sonnet 5 better for coding?
Claude Sonnet 5 scores higher on coding benchmarks: 55.5 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Claude Sonnet 4.6 and Claude Sonnet 5 share?
44 benchmarks have published results for both models. Claude Sonnet 4.6 has 57 scored results on Noometry and Claude Sonnet 5 has 51.