Model comparison
GPT-5.6 Sol vs GPT-6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 61.8 on the Noometry Index. GPT-6 Sol 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 . 45 shared benchmarks.
Summary
- They share 45 benchmarks with published results for both. GPT-5.6 Sol scores higher in 8 categories and GPT-6 Sol in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5.6 Sol leads 50.3 to 37.2.
- The biggest single-benchmark swing is SimpleQA Verified: 69.7% for GPT-5.6 Sol and 60.7% for GPT-6 Sol.
- GPT-6 Sol is cheaper at $2 / $10 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
Side by side
| GPT-5.6 Sol | GPT-6 Sol | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 65.0 | 61.8 |
| Released | 2026-07-09 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $4 | $2 |
| Output $ / M tokens | $20 | $10 |
| Results tracked | 65 | 45 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GPT-5.6 Sol: 65.1 (#7), GPT-6 Sol: 60.1 (#11)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| DeepSWE | 72.7% | 68.8% |
| FrontierCode | 47.5% | 49.3% |
| LMArena WebDev | 1618 | 1688 |
| SciCode | 57.1% | 57.6% |
| LMArena Coding | 1498 | 1447 |
| ALE-Bench | 2,177 | 2,462 |
| CursorBench | 41.7% | — |
| FrontierSWE | 32.2% | — |
| GSO | 76.5% | — |
| WeirdML | 89.4% | — |
| MirrorCode | 20% | — |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-5.6 Sol: 50.3 (#7), GPT-6 Sol: 37.2 (#36)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| APEX-Agents | 51.4% | 54.3% |
| GDP.pdf | 30.7% | 26.4% |
| Vending-Bench 2 | 9,619 | 14,428 |
| OSWorld 2.0 | 27.3% | — |
| τ²-bench Banking | 46.9% | — |
| PostTrainBench | 36.2% | — |
| BALROG | 60% | — |
| GBAEval | 52.6% | — |
| LMArena Search | 1257 | — |
Reasoning Too close to call
GPT-5.6 Sol: 74.8 (#8), GPT-6 Sol: 74.0 (#9)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| ARC-AGI-2 | 92.5% | 89.6% |
| NYT Connections (extended) | 93.8% | 90.1% |
| ARC-AGI-1 | 97.5% | 95.5% |
| CritPt | 32.3% | 30.9% |
| EBR-Bench | 44.8% | 53.3% |
| LMArena Hard Prompts | 1484 | 1418 |
| Mystery Game Puzzles | 58% | 56% |
| DTBench | 96% | 97.3% |
| LMCA | 59.2% | 59.1% |
| Epoch Capabilities Index | 161.66 | 162.72 |
| SimpleBench | 71.7% | — |
| Kagi LLM Benchmark | 67% | — |
| Chess Puzzles | 64% | — |
| EnigmaEval | 37.1% | — |
| Surface Evolver Bench | 93.1% | — |
| Bench to the Future 3 | 0.14 | — |
Math GPT-6 Sol leads
GPT-5.6 Sol: 85.6 (#9), GPT-6 Sol: 87.2 (#7)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 89.1% | 89.8% |
| FrontierMath Tier 4 | 82.9% | 90% |
| OTIS Mock AIME 2024-2025 | 100% | 100% |
| ProofBench | 83% | 83% |
| LMArena Math | 1474 | 1402 |
| FrontierMath Erdős | 0% | — |
Knowledge Too close to call
GPT-5.6 Sol: 64.3 (#18), GPT-6 Sol: 64.8 (#15)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| GPQA Diamond | 93.5% | 94.3% |
| SimpleQA Verified | 69.7% | 60.7% |
| Vectara Hallucination Rate | 12.4% | 6.5% |
| LMArena Expert | 1516 | 1439 |
Multimodal GPT-5.6 Sol leads
GPT-5.6 Sol: 48.6 (#9), GPT-6 Sol: 47.6 (#10)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| LMArena Vision | 1281 | 1245 |
| Blueprint-Bench 2 | 33.6% | 36.9% |
| Furniture Assembly | 56.7% | 58.3% |
| LMArena Document | 1483 | — |
Multilingual GPT-5.6 Sol leads
GPT-5.6 Sol: 55.3 (#32), GPT-6 Sol: 50.5 (#118)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| LMArena Non-English | 1452 | 1385 |
| LMArena Chinese | 1527 | 1405 |
| LMArena French | 1477 | 1410 |
| LMArena German | 1476 | 1390 |
| LMArena Japanese | 1471 | 1385 |
| LMArena Korean | 1442 | 1341 |
| LMArena Russian | 1468 | 1401 |
| LMArena Spanish | 1441 | 1384 |
Instruction Following GPT-5.6 Sol leads
GPT-5.6 Sol: 77.7 (#16), GPT-6 Sol: 74.5 (#94)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| LMArena Instruction Following | 1482 | 1412 |
Long Context GPT-5.6 Sol leads
GPT-5.6 Sol: 45.4 (#42), GPT-6 Sol: 43.1 (#108)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| LMArena Longer Query | 1480 | 1411 |
Writing & Preference GPT-5.6 Sol leads
GPT-5.6 Sol: 73.3 (#12), GPT-6 Sol: 71.9 (#18)
| Benchmark | GPT-5.6 Sol | GPT-6 Sol |
|---|---|---|
| LMArena Text | 1457 | 1395 |
| LMArena Creative Writing | 1448 | 1378 |
| EQ-Bench Creative Writing | 1972 | 2125 |
| LMArena Multi-Turn | 1460 | 1412 |
| EQ-Bench 4 | 1250 | — |
Frequently asked questions
Is GPT-5.6 Sol better than GPT-6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 61.8 on the Noometry Index. GPT-6 Sol 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, GPT-5.6 Sol or GPT-6 Sol?
GPT-6 Sol is cheaper. It lists at $2 per million input tokens and $10 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GPT-5.6 Sol or GPT-6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 60.1 in the Noometry coding category.
Which has the bigger context window?
Both accept 1.05M tokens.
How many benchmarks do GPT-5.6 Sol and GPT-6 Sol share?
45 benchmarks have published results for both models. GPT-5.6 Sol has 65 scored results on Noometry and GPT-6 Sol has 45.