Model comparison
Claude Sonnet 5 vs GPT-5.6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 54.6 on the Noometry Index. Claude Sonnet 5 costs 2.0× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 50 shared benchmarks.
Summary
- They share 50 benchmarks with published results for both. Claude Sonnet 5 scores higher in 0 categories and GPT-5.6 Sol in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 49.1.
- The biggest single-benchmark swing is FrontierMath Tier 4: 29.3% for Claude Sonnet 5 and 82.9% for GPT-5.6 Sol.
- Claude Sonnet 5 is cheaper at $2 / $10 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Sonnet 5 | GPT-5.6 Sol | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 54.6 | 65.0 |
| Released | 2026-06-29 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $2 | $4 |
| Output $ / M tokens | $10 | $20 |
| Results tracked | 51 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
Claude Sonnet 5: 55.5 (#26), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| DeepSWE | 53.8% | 72.7% |
| FrontierCode | 42.7% | 47.5% |
| CursorBench | 34.1% | 41.7% |
| LMArena WebDev | 1541 | 1618 |
| SciCode | 54.3% | 57.1% |
| GSO | 37.3% | 76.5% |
| WeirdML | 68.8% | 89.4% |
| LMArena Coding | 1483 | 1498 |
| ALE-Bench | 1,463 | 2,177 |
| FrontierSWE | — | 32.2% |
| MirrorCode | — | 20% |
Agentic & Tool Use GPT-5.6 Sol leads
Claude Sonnet 5: 42.8 (#18), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| APEX-Agents | 54.5% | 51.4% |
| GBAEval | 65.3% | 52.6% |
| LMArena Search | 1194 | 1257 |
| Vending-Bench 2 | 6,378 | 9,619 |
| OSWorld 2.0 | — | 27.3% |
| τ²-bench Banking | — | 46.9% |
| PostTrainBench | — | 36.2% |
| BALROG | — | 60% |
| GDP.pdf | — | 30.7% |
Reasoning GPT-5.6 Sol leads
Claude Sonnet 5: 49.1 (#39), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| SimpleBench | 60.6% | 71.7% |
| NYT Connections (extended) | 75.1% | 93.8% |
| CritPt | 16.9% | 32.3% |
| Chess Puzzles | 35% | 64% |
| LMArena Hard Prompts | 1461 | 1484 |
| Mystery Game Puzzles | 35% | 58% |
| DTBench | 92.5% | 96% |
| LMCA | 50% | 59.2% |
| Surface Evolver Bench | 60% | 93.1% |
| Bench to the Future 3 | 0.14 | 0.14 |
| Epoch Capabilities Index | 156.21 | 161.66 |
| ARC-AGI-2 | — | 92.5% |
| Kagi LLM Benchmark | — | 67% |
| ARC-AGI-1 | — | 97.5% |
| EnigmaEval | — | 37.1% |
| EBR-Bench | — | 44.8% |
| ForecastBench | 61.1 | — |
Math GPT-5.6 Sol leads
Claude Sonnet 5: 66.2 (#27), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 65.6% | 89.1% |
| FrontierMath Tier 4 | 29.3% | 82.9% |
| OTIS Mock AIME 2024-2025 | 94.7% | 100% |
| ProofBench | 77% | 83% |
| LMArena Math | 1467 | 1474 |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.6 Sol leads
Claude Sonnet 5: 55.6 (#47), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 90.5% | 93.5% |
| SimpleQA Verified | 33.7% | 69.7% |
| LMArena Expert | 1490 | 1516 |
| Vectara Hallucination Rate | — | 12.4% |
Multimodal GPT-5.6 Sol leads
Claude Sonnet 5: 42.4 (#31), GPT-5.6 Sol: 48.6 (#9)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | 1274 | 1281 |
| Blueprint-Bench 2 | 24.9% | 33.6% |
| LMArena Document | 1466 | 1483 |
| Furniture Assembly | — | 56.7% |
Multilingual GPT-5.6 Sol leads
Claude Sonnet 5: 53.8 (#55), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1431 | 1452 |
| LMArena Chinese | 1477 | 1527 |
| LMArena French | 1460 | 1477 |
| LMArena German | 1440 | 1476 |
| LMArena Japanese | 1422 | 1471 |
| LMArena Korean | 1411 | 1442 |
| LMArena Russian | 1451 | 1468 |
| LMArena Spanish | 1437 | 1441 |
Instruction Following GPT-5.6 Sol leads
Claude Sonnet 5: 76.3 (#41), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1452 | 1482 |
Long Context Too close to call
Claude Sonnet 5: 44.8 (#55), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1463 | 1480 |
Writing & Preference GPT-5.6 Sol leads
Claude Sonnet 5: 69.2 (#25), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | Claude Sonnet 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1442 | 1457 |
| LMArena Creative Writing | 1416 | 1448 |
| EQ-Bench Creative Writing | 1794 | 1972 |
| EQ-Bench 4 | 1236 | 1250 |
| LMArena Multi-Turn | 1454 | 1460 |
Frequently asked questions
Is Claude Sonnet 5 better than GPT-5.6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 54.6 on the Noometry Index. Claude Sonnet 5 costs 2.0× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 5 or GPT-5.6 Sol?
Claude Sonnet 5 is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is Claude Sonnet 5 or GPT-5.6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 55.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 1M.
How many benchmarks do Claude Sonnet 5 and GPT-5.6 Sol share?
50 benchmarks have published results for both models. Claude Sonnet 5 has 51 scored results on Noometry and GPT-5.6 Sol has 65.