Model comparison
GPT-5 vs GPT-6 Sol
GPT-6 Sol is the stronger model overall, scoring 61.8 to 50.9 on the Noometry Index.
Last verified . 37 shared benchmarks.
Summary
- They share 37 benchmarks with published results for both. GPT-5 scores higher in 2 categories and GPT-6 Sol in 8 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Sol leads 74.0 to 38.3.
- The biggest single-benchmark swing is ARC-AGI-2: 9.9% for GPT-5 and 89.6% for GPT-6 Sol.
- GPT-5 is cheaper at $1.25 / $10 per million input/output tokens, against $2 / $10 for GPT-6 Sol.
- GPT-6 Sol accepts more context: 1.05M tokens versus 400K.
Side by side
| GPT-5 | GPT-6 Sol | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 50.9 | 61.8 |
| Released | 2025-08-07 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $1.25 | $2 |
| Output $ / M tokens | $10 | $10 |
| Results tracked | 69 | 45 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Sol leads
GPT-5: 50.3 (#47), GPT-6 Sol: 60.1 (#11)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena WebDev | 1418 | 1688 |
| SciCode | 42.9% | 57.6% |
| LMArena Coding | 1436 | 1447 |
| ALE-Bench | 1,162 | 2,462 |
| SWE-bench Verified | 73.6% | — |
| DeepSWE | — | 68.8% |
| FrontierCode | — | 49.3% |
| SWE-bench Verified (bash only) | 65% | — |
| Aider Polyglot | 88% | — |
| GSO | 6.9% | — |
| WeirdML | 60.7% | — |
| AlgoTune | 1.67 | — |
Agentic & Tool Use GPT-6 Sol leads
GPT-5: 33.1 (#56), GPT-6 Sol: 37.2 (#36)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| Terminal-Bench | 49.6% | — |
| APEX-Agents | — | 54.3% |
| GDPval | 34.8% | — |
| Remote Labor Index | 1.7% | — |
| DeepResearch Bench | 49.6% | — |
| BALROG | 32.8% | — |
| GDP.pdf | — | 26.4% |
| LMArena Search | 1133 | — |
| METR Time Horizons | 69.6% | — |
| Vending-Bench 2 | — | 14,428 |
Reasoning GPT-6 Sol leads
GPT-5: 38.3 (#64), GPT-6 Sol: 74.0 (#9)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| ARC-AGI-2 | 9.9% | 89.6% |
| ARC-AGI-1 | 65.7% | 95.5% |
| CritPt | 12.6% | 30.9% |
| EBR-Bench | 12.7% | 53.3% |
| LMArena Hard Prompts | 1416 | 1418 |
| Mystery Game Puzzles | 23% | 56% |
| DTBench | 90.7% | 97.3% |
| LMCA | 40% | 59.1% |
| Epoch Capabilities Index | 150 | 162.72 |
| SimpleBench | 56.7% | — |
| Kagi LLM Benchmark | 72.7% | — |
| NYT Connections (extended) | — | 90.1% |
| Chess Puzzles | 37% | — |
| EnigmaEval | 10.5% | — |
| ForecastBench | 61.4 | — |
Math GPT-6 Sol leads
GPT-5: 55.0 (#44), GPT-6 Sol: 87.2 (#7)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 55.4% | 89.8% |
| FrontierMath Tier 4 | 22% | 90% |
| OTIS Mock AIME 2024-2025 | 91.4% | 100% |
| ProofBench | 18% | 83% |
| LMArena Math | 1407 | 1402 |
| Omni-MATH | 64.7% | — |
| MATH Level 5 | 98.1% | — |
| FrontierMath (Feb 2025 set) | 32.4% | — |
| FrontierMath Tier 4 (v1) | 12.5% | — |
Knowledge GPT-6 Sol leads
GPT-5: 56.6 (#43), GPT-6 Sol: 64.8 (#15)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| GPQA Diamond | 86.2% | 94.3% |
| SimpleQA Verified | 50.1% | 60.7% |
| Vectara Hallucination Rate | 14.7% | 6.5% |
| LMArena Expert | 1419 | 1439 |
| Humanity's Last Exam | 25.3% | — |
| MMLU-Pro | 86.3% | — |
| Confabulations | 10.3% | — |
| GPQA (HELM) | 79.2% | — |
Multimodal Too close to call
GPT-5: 46.8 (#13), GPT-6 Sol: 47.6 (#10)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena Vision | 1232 | 1245 |
| GeoBench | 81% | — |
| VPCT | 66% | — |
| Blueprint-Bench 2 | — | 36.9% |
| Furniture Assembly | — | 58.3% |
Multilingual Too close to call
GPT-5: 51.4 (#110), GPT-6 Sol: 50.5 (#118)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena Non-English | 1397 | 1385 |
| LMArena Chinese | 1422 | 1405 |
| LMArena French | 1410 | 1410 |
| LMArena German | 1416 | 1390 |
| LMArena Japanese | 1409 | 1385 |
| LMArena Korean | 1360 | 1341 |
| LMArena Russian | 1406 | 1401 |
| LMArena Spanish | 1399 | 1384 |
Instruction Following Too close to call
GPT-5: 73.8 (#113), GPT-6 Sol: 74.5 (#94)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena Instruction Following | 1388 | 1412 |
| IFEval | 87.5% | — |
Long Context GPT-5 leads
GPT-5: 69.5 (#2), GPT-6 Sol: 43.1 (#108)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena Longer Query | 1399 | 1411 |
| Fiction.LiveBench | 97.2% | — |
Writing & Preference GPT-6 Sol leads
GPT-5: 63.4 (#65), GPT-6 Sol: 71.9 (#18)
| Benchmark | GPT-5 | GPT-6 Sol |
|---|---|---|
| LMArena Text | 1406 | 1395 |
| LMArena Creative Writing | 1365 | 1378 |
| EQ-Bench Creative Writing | 1627 | 2125 |
| LMArena Multi-Turn | 1426 | 1412 |
| Short-Story Creative Writing | 86% | — |
| WildBench | 85.7% | — |
Frequently asked questions
Is GPT-5 better than GPT-6 Sol?
GPT-6 Sol is the stronger model overall, scoring 61.8 to 50.9 on the Noometry Index.
Which is cheaper, GPT-5 or GPT-6 Sol?
GPT-5 is cheaper. It lists at $1.25 per million input tokens and $10 per million output tokens; GPT-6 Sol lists at $2 and $10.
Is GPT-5 or GPT-6 Sol better for coding?
GPT-6 Sol scores higher on coding benchmarks: 60.1 versus 50.3 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Sol does, with 1.05M tokens against 400K.
How many benchmarks do GPT-5 and GPT-6 Sol share?
37 benchmarks have published results for both models. GPT-5 has 69 scored results on Noometry and GPT-6 Sol has 45.