Model comparison
GPT-5.6 Sol vs GPT-5-Codex
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 37.9 on the Noometry Index. GPT-5-Codex costs 2.3× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 2 shared benchmarks.
Summary
- They share 2 benchmarks with published results for both. GPT-5.6 Sol scores higher in 3 categories and GPT-5-Codex in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.6 Sol leads 74.8 to 30.9.
- The biggest single-benchmark swing is WeirdML: 89.4% for GPT-5.6 Sol and 54.5% for GPT-5-Codex.
- GPT-5-Codex is cheaper at $1.25 / $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 400K.
Side by side
| GPT-5.6 Sol | GPT-5-Codex | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 65.0 | 37.9 |
| Released | 2026-07-09 | 2025-09-15 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 400K |
| Max output | 128K | 128K |
| Input $ / M tokens | $4 | $1.25 |
| Output $ / M tokens | $20 | $10 |
| Results tracked | 65 | 3 |
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-5-Codex: 42.4 (#103)
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| WeirdML | 89.4% | 54.5% |
| DeepSWE | 72.7% | — |
| FrontierCode | 47.5% | — |
| CursorBench | 41.7% | — |
| LMArena WebDev | 1618 | — |
| FrontierSWE | 32.2% | — |
| SciCode | 57.1% | — |
| GSO | 76.5% | — |
| LMArena Coding | 1498 | — |
| MirrorCode | 20% | — |
| ALE-Bench | 2,177 | — |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-5.6 Sol: 50.3 (#7), GPT-5-Codex: 31.0 (#72)
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| Terminal-Bench | — | 44.3% |
| APEX-Agents | 51.4% | — |
| OSWorld 2.0 | 27.3% | — |
| τ²-bench Banking | 46.9% | — |
| PostTrainBench | 36.2% | — |
| BALROG | 60% | — |
| GBAEval | 52.6% | — |
| GDP.pdf | 30.7% | — |
| LMArena Search | 1257 | — |
| Vending-Bench 2 | 9,619 | — |
Reasoning GPT-5.6 Sol leads
GPT-5.6 Sol: 74.8 (#8), GPT-5-Codex: 30.9 (#83)
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| Kagi LLM Benchmark | 67% | 70.3% |
| ARC-AGI-2 | 92.5% | — |
| SimpleBench | 71.7% | — |
| NYT Connections (extended) | 93.8% | — |
| ARC-AGI-1 | 97.5% | — |
| CritPt | 32.3% | — |
| Chess Puzzles | 64% | — |
| EnigmaEval | 37.1% | — |
| EBR-Bench | 44.8% | — |
| LMArena Hard Prompts | 1484 | — |
| Mystery Game Puzzles | 58% | — |
| DTBench | 96% | — |
| LMCA | 59.2% | — |
| Surface Evolver Bench | 93.1% | — |
| Bench to the Future 3 | 0.14 | — |
| Epoch Capabilities Index | 161.66 | — |
Math Not comparable
GPT-5.6 Sol: 85.6 (#9), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| FrontierMath (Tiers 1-3) | 89.1% | — |
| FrontierMath Tier 4 | 82.9% | — |
| OTIS Mock AIME 2024-2025 | 100% | — |
| ProofBench | 83% | — |
| LMArena Math | 1474 | — |
| FrontierMath Erdős | 0% | — |
Knowledge Not comparable
GPT-5.6 Sol: 64.3 (#18), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| GPQA Diamond | 93.5% | — |
| SimpleQA Verified | 69.7% | — |
| Vectara Hallucination Rate | 12.4% | — |
| LMArena Expert | 1516 | — |
Multimodal Not comparable
GPT-5.6 Sol: 48.6 (#9), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| LMArena Vision | 1281 | — |
| Blueprint-Bench 2 | 33.6% | — |
| Furniture Assembly | 56.7% | — |
| LMArena Document | 1483 | — |
Multilingual Not comparable
GPT-5.6 Sol: 55.3 (#32), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| LMArena Non-English | 1452 | — |
| LMArena Chinese | 1527 | — |
| LMArena French | 1477 | — |
| LMArena German | 1476 | — |
| LMArena Japanese | 1471 | — |
| LMArena Korean | 1442 | — |
| LMArena Russian | 1468 | — |
| LMArena Spanish | 1441 | — |
Instruction Following Not comparable
GPT-5.6 Sol: 77.7 (#16), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| LMArena Instruction Following | 1482 | — |
Long Context Not comparable
GPT-5.6 Sol: 45.4 (#42), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| LMArena Longer Query | 1480 | — |
Writing & Preference Not comparable
GPT-5.6 Sol: 73.3 (#12), GPT-5-Codex: —
| Benchmark | GPT-5.6 Sol | GPT-5-Codex |
|---|---|---|
| LMArena Text | 1457 | — |
| LMArena Creative Writing | 1448 | — |
| EQ-Bench Creative Writing | 1972 | — |
| EQ-Bench 4 | 1250 | — |
| LMArena Multi-Turn | 1460 | — |
Frequently asked questions
Is GPT-5.6 Sol better than GPT-5-Codex?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 37.9 on the Noometry Index. GPT-5-Codex costs 2.3× 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-5-Codex?
GPT-5-Codex is cheaper. It lists at $1.25 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-5-Codex better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 42.4 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 400K.
How many benchmarks do GPT-5.6 Sol and GPT-5-Codex share?
2 benchmarks have published results for both models. GPT-5.6 Sol has 65 scored results on Noometry and GPT-5-Codex has 3.