Model comparison
GPT-5.5 vs GPT-6 Sol
GPT-5.5 is the stronger model overall, scoring 63.4 to 61.8 on the Noometry Index. GPT-6 Sol costs 2.8× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Last verified . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. GPT-5.5 scores higher in 5 categories and GPT-6 Sol in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5.5 leads 50.7 to 37.2.
- The biggest single-benchmark swing is ProofBench: 50% for GPT-5.5 and 83% for GPT-6 Sol.
- GPT-6 Sol is cheaper at $2 / $10 per million input/output tokens, against $5 / $30 for GPT-5.5.
Side by side
| GPT-5.5 | GPT-6 Sol | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 63.4 | 61.8 |
| Released | 2026-04-23 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $5 | $2 |
| Output $ / M tokens | $30 | $10 |
| Results tracked | 71 | 45 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Sol leads
GPT-5.5: 58.2 (#17), GPT-6 Sol: 60.1 (#11)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| DeepSWE | 67% | 68.8% |
| FrontierCode | 43% | 49.3% |
| LMArena WebDev | 1513 | 1688 |
| SciCode | 56.1% | 57.6% |
| LMArena Coding | 1494 | 1447 |
| ALE-Bench | 1,943 | 2,462 |
| SWE-bench Verified | 80.6% | — |
| GSO | 40.2% | — |
| WeirdML | 84.9% | — |
| MirrorCode | 10% | — |
Agentic & Tool Use GPT-5.5 leads
GPT-5.5: 50.7 (#6), GPT-6 Sol: 37.2 (#36)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| APEX-Agents | 55.1% | 54.3% |
| GDP.pdf | 26% | 26.4% |
| Vending-Bench 2 | 7,524 | 14,428 |
| Terminal-Bench | 84.7% | — |
| OSWorld 2.0 | 13% | — |
| Remote Labor Index | 6.3% | — |
| τ²-bench Banking | 44.6% | — |
| DeepResearch Bench | 54% | — |
| PostTrainBench | 27.2% | — |
| ExploitBench | 47.4% | — |
| GBAEval | 53.2% | — |
| LMArena Search | 1242 | — |
Reasoning GPT-6 Sol leads
GPT-5.5: 72.8 (#11), GPT-6 Sol: 74.0 (#9)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| ARC-AGI-2 | 85% | 89.6% |
| NYT Connections (extended) | 96.2% | 90.1% |
| ARC-AGI-1 | 95% | 95.5% |
| CritPt | 27.1% | 30.9% |
| EBR-Bench | 34.3% | 53.3% |
| LMArena Hard Prompts | 1489 | 1418 |
| Mystery Game Puzzles | 56% | 56% |
| DTBench | 96% | 97.3% |
| LMCA | 54.3% | 59.1% |
| Epoch Capabilities Index | 159.1 | 162.72 |
| SimpleBench | 69% | — |
| Kagi LLM Benchmark | 88.8% | — |
| Chess Puzzles | 54% | — |
| Surface Evolver Bench | 88.1% | — |
| Bench to the Future 3 | 0.14 | — |
| ForecastBench | 60.6 | — |
Math GPT-6 Sol leads
GPT-5.5: 81.7 (#11), GPT-6 Sol: 87.2 (#7)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 85.3% | 89.8% |
| FrontierMath Tier 4 | 72.5% | 90% |
| OTIS Mock AIME 2024-2025 | 100% | 100% |
| ProofBench | 50% | 83% |
| LMArena Math | 1486 | 1402 |
| MathArena Final-Answer Competitions | 94.3% | — |
| FrontierMath (Feb 2025 set) | 51.7% | — |
| FrontierMath Erdős | 0% | — |
| FrontierMath Tier 4 (v1) | 35.4% | — |
Knowledge Too close to call
GPT-5.5: 64.4 (#17), GPT-6 Sol: 64.8 (#15)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| GPQA Diamond | 94% | 94.3% |
| SimpleQA Verified | 63% | 60.7% |
| Vectara Hallucination Rate | 9.3% | 6.5% |
| LMArena Expert | 1508 | 1439 |
Multimodal Too close to call
GPT-5.5: 46.9 (#12), GPT-6 Sol: 47.6 (#10)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| LMArena Vision | 1297 | 1245 |
| Blueprint-Bench 2 | 36.2% | 36.9% |
| Furniture Assembly | 44.2% | 58.3% |
| LMArena Document | 1486 | — |
Multilingual GPT-5.5 leads
GPT-5.5: 56.4 (#20), GPT-6 Sol: 50.5 (#118)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| LMArena Non-English | 1467 | 1385 |
| LMArena Chinese | 1533 | 1405 |
| LMArena French | 1486 | 1410 |
| LMArena German | 1480 | 1390 |
| LMArena Japanese | 1498 | 1385 |
| LMArena Korean | 1460 | 1341 |
| LMArena Russian | 1473 | 1401 |
| LMArena Spanish | 1468 | 1384 |
Instruction Following GPT-5.5 leads
GPT-5.5: 77.5 (#18), GPT-6 Sol: 74.5 (#94)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| LMArena Instruction Following | 1479 | 1412 |
Long Context GPT-5.5 leads
GPT-5.5: 48.3 (#12), GPT-6 Sol: 43.1 (#108)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| LMArena Longer Query | 1484 | 1411 |
| CL-bench Life | 22.2% | — |
Writing & Preference Too close to call
GPT-5.5: 72.7 (#13), GPT-6 Sol: 71.9 (#18)
| Benchmark | GPT-5.5 | GPT-6 Sol |
|---|---|---|
| LMArena Text | 1472 | 1395 |
| LMArena Creative Writing | 1455 | 1378 |
| EQ-Bench Creative Writing | 1844 | 2125 |
| LMArena Multi-Turn | 1476 | 1412 |
| EQ-Bench 4 | 1315 | — |
Frequently asked questions
Is GPT-5.5 better than GPT-6 Sol?
GPT-5.5 is the stronger model overall, scoring 63.4 to 61.8 on the Noometry Index. GPT-6 Sol costs 2.8× less per token, which makes it the better buy when GPT-5.5's lead doesn't matter for your workload.
Which is cheaper, GPT-5.5 or GPT-6 Sol?
GPT-6 Sol is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; GPT-5.5 lists at $5 and $30.
Is GPT-5.5 or GPT-6 Sol better for coding?
GPT-6 Sol scores higher on coding benchmarks: 60.1 versus 58.2 in the Noometry coding category.
Which has the bigger context window?
Both accept 1.05M tokens.
How many benchmarks do GPT-5.5 and GPT-6 Sol share?
45 benchmarks have published results for both models. GPT-5.5 has 71 scored results on Noometry and GPT-6 Sol has 45.