Model comparison
GPT-4o vs GPT-5.6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 28.6 on the Noometry Index. GPT-4o costs 1.8× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. GPT-4o scores higher in 0 categories and GPT-5.6 Sol in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.6 Sol leads 85.6 to 10.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 6.4% for GPT-4o and 100% for GPT-5.6 Sol.
- GPT-4o is cheaper at $2.50 / $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 128K.
Side by side
| GPT-4o | GPT-5.6 Sol | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 28.6 | 65.0 |
| Released | 2024-05-13 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 128K | 1.05M |
| Max output | 16K | 128K |
| Input $ / M tokens | $2.50 | $4 |
| Output $ / M tokens | $10 | $20 |
| Results tracked | 72 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GPT-4o: 24.8 (#328), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| GSO | 0% | 76.5% |
| WeirdML | 25.1% | 89.4% |
| LMArena Coding | 1297 | 1498 |
| SWE-bench Verified | 31% | — |
| DeepSWE | — | 72.7% |
| FrontierCode | — | 47.5% |
| SWE-bench Verified (bash only) | 21.6% | — |
| Aider Polyglot | 45.3% | — |
| CursorBench | — | 41.7% |
| LMArena WebDev | — | 1618 |
| FrontierSWE | — | 32.2% |
| SciCode | — | 57.1% |
| BigCodeBench Instruct | 51.1% | — |
| LiveBench Coding | 51.4% | — |
| MirrorCode | — | 20% |
| BigCodeBench Complete | 61.1% | — |
| CadEval | 26% | — |
| ALE-Bench | — | 2,177 |
| HumanEval+ | 87.2% | — |
| MBPP+ | 72.2% | — |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-4o: 21.0 (#141), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| BALROG | 32.3% | 60% |
| LMArena Search | 1006 | 1257 |
| APEX-Agents | — | 51.4% |
| OSWorld 2.0 | — | 27.3% |
| GDPval | 9.9% | — |
| TheAgentCompany | 8.6% | — |
| τ²-bench Banking | — | 46.9% |
| Cybench | 12.5% | — |
| PostTrainBench | — | 36.2% |
| GBAEval | — | 52.6% |
| GDP.pdf | — | 30.7% |
| METR Time Horizons | 40.8% | — |
| Vending-Bench 2 | — | 9,619 |
Reasoning GPT-5.6 Sol leads
GPT-4o: 9.4 (#343), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-2 | 0% | 92.5% |
| SimpleBench | 17.8% | 71.7% |
| ARC-AGI-1 | 4.5% | 97.5% |
| CritPt | 0% | 32.3% |
| Chess Puzzles | 13% | 64% |
| EnigmaEval | 0.8% | 37.1% |
| LMArena Hard Prompts | 1281 | 1484 |
| DTBench | 64.5% | 96% |
| LMCA | 16.6% | 59.2% |
| Epoch Capabilities Index | 128.97 | 161.66 |
| Kagi LLM Benchmark | — | 67% |
| NYT Connections (extended) | — | 93.8% |
| EBR-Bench | — | 44.8% |
| LiveBench Reasoning | 55.8% | — |
| Mystery Game Puzzles | — | 58% |
| LiveBench Data Analysis | 60.9% | — |
| Surface Evolver Bench | — | 93.1% |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 57.7 | — |
| LiveBench | 55.3% | — |
Math GPT-5.6 Sol leads
GPT-4o: 10.6 (#312), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 0.4% | 89.1% |
| OTIS Mock AIME 2024-2025 | 6.4% | 100% |
| LMArena Math | 1285 | 1474 |
| FrontierMath Tier 4 | — | 82.9% |
| ProofBench | — | 83% |
| Omni-MATH | 29.3% | — |
| LiveBench Math | 49.5% | — |
| MATH Level 5 | 53.3% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.6 Sol leads
GPT-4o: 28.8 (#242), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 49.2% | 93.5% |
| SimpleQA Verified | 26% | 69.7% |
| Vectara Hallucination Rate | 9.6% | 12.4% |
| LMArena Expert | 1250 | 1516 |
| Humanity's Last Exam | 2.7% | — |
| MMLU-Pro | 71.3% | — |
| Confabulations | 15.3% | — |
| GPQA (HELM) | 52% | — |
| MMLU | 88.1% | — |
Multimodal GPT-5.6 Sol leads
GPT-4o: 34.5 (#91), GPT-5.6 Sol: 48.6 (#9)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | 1137 | 1281 |
| Video-MME | 71.9% | — |
| GeoBench | 71% | — |
| VPCT | 40% | — |
| Blueprint-Bench 2 | — | 33.6% |
| Furniture Assembly | — | 56.7% |
| LMArena Document | — | 1483 |
| ScienceQA | 88.5% | — |
Multilingual GPT-5.6 Sol leads
GPT-4o: 43.2 (#186), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1283 | 1452 |
| LMArena Chinese | 1277 | 1527 |
| LMArena French | 1304 | 1477 |
| LMArena German | 1282 | 1476 |
| LMArena Japanese | 1257 | 1471 |
| LMArena Korean | 1234 | 1442 |
| LMArena Russian | 1286 | 1468 |
| LMArena Spanish | 1292 | 1441 |
Instruction Following GPT-5.6 Sol leads
GPT-4o: 66.6 (#207), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1278 | 1482 |
| LiveBench Instruction Following | 68.6% | — |
| IFEval | 81.7% | — |
Long Context GPT-5.6 Sol leads
GPT-4o: 39.4 (#179), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1289 | 1480 |
| Fiction.LiveBench | 66.7% | — |
Writing & Preference GPT-5.6 Sol leads
GPT-4o: 52.6 (#166), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | GPT-4o | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1300 | 1457 |
| LMArena Creative Writing | 1292 | 1448 |
| LMArena Multi-Turn | 1302 | 1460 |
| Short-Story Creative Writing | 81.8% | — |
| EQ-Bench Creative Writing | — | 1972 |
| WildBench | 82.8% | — |
| EQ-Bench 4 | — | 1250 |
| LiveBench Language | 47.6% | — |
Frequently asked questions
Is GPT-4o better than GPT-5.6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 28.6 on the Noometry Index. GPT-4o costs 1.8× 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-4o or GPT-5.6 Sol?
GPT-4o is cheaper. It lists at $2.50 per million input tokens and $10 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GPT-4o or GPT-5.6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 24.8 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 128K.
How many benchmarks do GPT-4o and GPT-5.6 Sol share?
36 benchmarks have published results for both models. GPT-4o has 72 scored results on Noometry and GPT-5.6 Sol has 65.