Model comparison
Claude Sonnet 4.6 vs GPT-5.5
GPT-5.5 is the stronger model overall, scoring 63.4 to 50.3 on the Noometry Index. Claude Sonnet 4.6 costs 1.9× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Last verified . 55 shared benchmarks.
Summary
- They share 55 benchmarks with published results for both. Claude Sonnet 4.6 scores higher in 0 categories and GPT-5.5 in 10 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.5 leads 81.7 to 52.9.
- The biggest single-benchmark swing is Chess Puzzles: 13% for Claude Sonnet 4.6 and 54% for GPT-5.5.
- Claude Sonnet 4.6 is cheaper at $3 / $15 per million input/output tokens, against $5 / $30 for GPT-5.5.
- GPT-5.5 accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Sonnet 4.6 | GPT-5.5 | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 50.3 | 63.4 |
| Released | 2026-02-17 | 2026-04-23 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $3 | $5 |
| Output $ / M tokens | $15 | $30 |
| Results tracked | 57 | 71 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.5 leads
Claude Sonnet 4.6: 46.3 (#67), GPT-5.5: 58.2 (#17)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| SWE-bench Verified | 75.2% | 80.6% |
| DeepSWE | 29.9% | 67% |
| FrontierCode | 24.3% | 43% |
| LMArena WebDev | 1522 | 1513 |
| SciCode | 46.8% | 56.1% |
| WeirdML | 66.1% | 84.9% |
| LMArena Coding | 1504 | 1494 |
| ALE-Bench | 1,327 | 1,943 |
| GSO | — | 40.2% |
| MirrorCode | — | 10% |
Agentic & Tool Use GPT-5.5 leads
Claude Sonnet 4.6: 39.1 (#28), GPT-5.5: 50.7 (#6)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| Terminal-Bench | 53.4% | 84.7% |
| APEX-Agents | 43% | 55.1% |
| OSWorld 2.0 | 9.3% | 13% |
| DeepResearch Bench | 54.9% | 54% |
| ExploitBench | 23.6% | 47.4% |
| GBAEval | 48.8% | 53.2% |
| GDP.pdf | 18% | 26% |
| LMArena Search | 1221 | 1242 |
| Vending-Bench 2 | 7,204 | 7,524 |
| Remote Labor Index | — | 6.3% |
| τ²-bench Banking | — | 44.6% |
| OSWorld | 72.1% | — |
| PostTrainBench | — | 27.2% |
Reasoning GPT-5.5 leads
Claude Sonnet 4.6: 46.1 (#45), GPT-5.5: 72.8 (#11)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| ARC-AGI-2 | 60.4% | 85% |
| NYT Connections (extended) | 80.9% | 96.2% |
| ARC-AGI-1 | 86.5% | 95% |
| CritPt | 3.1% | 27.1% |
| Chess Puzzles | 13% | 54% |
| LMArena Hard Prompts | 1484 | 1489 |
| Mystery Game Puzzles | 16% | 56% |
| DTBench | 89.9% | 96% |
| LMCA | 46.5% | 54.3% |
| Epoch Capabilities Index | 152.24 | 159.1 |
| ForecastBench | 62 | 60.6 |
| SimpleBench | — | 69% |
| Kagi LLM Benchmark | — | 88.8% |
| Thematic Generalization | 76.3% | — |
| EBR-Bench | — | 34.3% |
| Surface Evolver Bench | — | 88.1% |
| Bench to the Future 3 | — | 0.14 |
Math GPT-5.5 leads
Claude Sonnet 4.6: 52.9 (#49), GPT-5.5: 81.7 (#11)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 85.8% | 100% |
| ProofBench | 45% | 50% |
| LMArena Math | 1462 | 1486 |
| FrontierMath (Feb 2025 set) | 32.4% | 51.7% |
| FrontierMath Tier 4 (v1) | 8.3% | 35.4% |
| FrontierMath (Tiers 1-3) | — | 85.3% |
| FrontierMath Tier 4 | — | 72.5% |
| MathArena Final-Answer Competitions | — | 94.3% |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.5 leads
Claude Sonnet 4.6: 51.7 (#65), GPT-5.5: 64.4 (#17)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| GPQA Diamond | 87.4% | 94% |
| SimpleQA Verified | 35.5% | 63% |
| Vectara Hallucination Rate | 10.6% | 9.3% |
| LMArena Expert | 1500 | 1508 |
Multimodal GPT-5.5 leads
Claude Sonnet 4.6: 38.0 (#68), GPT-5.5: 46.9 (#12)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| LMArena Vision | 1283 | 1297 |
| Blueprint-Bench 2 | 6.7% | 36.2% |
| LMArena Document | 1482 | 1486 |
| Furniture Assembly | — | 44.2% |
Multilingual GPT-5.5 leads
Claude Sonnet 4.6: 54.4 (#41), GPT-5.5: 56.4 (#20)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| LMArena Non-English | 1440 | 1467 |
| LMArena Chinese | 1491 | 1533 |
| LMArena French | 1465 | 1486 |
| LMArena German | 1428 | 1480 |
| LMArena Japanese | 1420 | 1498 |
| LMArena Korean | 1411 | 1460 |
| LMArena Russian | 1440 | 1473 |
| LMArena Spanish | 1464 | 1468 |
Instruction Following Too close to call
Claude Sonnet 4.6: 77.4 (#25), GPT-5.5: 77.5 (#18)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| LMArena Instruction Following | 1475 | 1479 |
Long Context GPT-5.5 leads
Claude Sonnet 4.6: 45.3 (#44), GPT-5.5: 48.3 (#12)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| LMArena Longer Query | 1479 | 1484 |
| CL-bench Life | — | 22.2% |
Writing & Preference GPT-5.5 leads
Claude Sonnet 4.6: 70.2 (#22), GPT-5.5: 72.7 (#13)
| Benchmark | Claude Sonnet 4.6 | GPT-5.5 |
|---|---|---|
| LMArena Text | 1458 | 1472 |
| LMArena Creative Writing | 1435 | 1455 |
| EQ-Bench Creative Writing | 1810 | 1844 |
| EQ-Bench 4 | 1207 | 1315 |
| LMArena Multi-Turn | 1464 | 1476 |
Frequently asked questions
Is Claude Sonnet 4.6 better than GPT-5.5?
GPT-5.5 is the stronger model overall, scoring 63.4 to 50.3 on the Noometry Index. Claude Sonnet 4.6 costs 1.9× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4.6 or GPT-5.5?
Claude Sonnet 4.6 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-5.5 lists at $5 and $30.
Is Claude Sonnet 4.6 or GPT-5.5 better for coding?
GPT-5.5 scores higher on coding benchmarks: 58.2 versus 46.3 in the Noometry coding category.
Which has the bigger context window?
GPT-5.5 does, with 1.05M tokens against 1M.
How many benchmarks do Claude Sonnet 4.6 and GPT-5.5 share?
55 benchmarks have published results for both models. Claude Sonnet 4.6 has 57 scored results on Noometry and GPT-5.5 has 71.