Model comparison
GPT-5.6 Luna vs GPT-5.6 Sol
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 18× less per token, which makes it the better buy when GPT-5.6 Sol's lead doesn't matter for your workload.
Last verified . 52 shared benchmarks.
Summary
- They share 52 benchmarks with published results for both. GPT-5.6 Luna 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 reasoning, where GPT-5.6 Sol leads 74.8 to 47.6.
- The biggest single-benchmark swing is Mystery Game Puzzles: 21% for GPT-5.6 Luna and 58% for GPT-5.6 Sol.
- GPT-5.6 Luna is cheaper at $0.20 / $1.20 per million input/output tokens, against $4 / $20 for GPT-5.6 Sol.
Side by side
| GPT-5.6 Luna | GPT-5.6 Sol | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 54.6 | 65.0 |
| Released | 2026-07-09 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.20 | $4 |
| Output $ / M tokens | $1.20 | $20 |
| Results tracked | 52 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.6 Sol leads
GPT-5.6 Luna: 54.5 (#28), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| DeepSWE | 67.2% | 72.7% |
| FrontierCode | 39.8% | 47.5% |
| CursorBench | 35.9% | 41.7% |
| LMArena WebDev | 1519 | 1618 |
| SciCode | 53.6% | 57.1% |
| WeirdML | 60.9% | 89.4% |
| LMArena Coding | 1466 | 1498 |
| ALE-Bench | 1,667 | 2,177 |
| FrontierSWE | — | 32.2% |
| GSO | — | 76.5% |
| MirrorCode | — | 20% |
Agentic & Tool Use GPT-5.6 Sol leads
GPT-5.6 Luna: 34.4 (#45), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| APEX-Agents | 43% | 51.4% |
| BALROG | 45.6% | 60% |
| GDP.pdf | 22.7% | 30.7% |
| Vending-Bench 2 | 4,095 | 9,619 |
| OSWorld 2.0 | — | 27.3% |
| τ²-bench Banking | — | 46.9% |
| PostTrainBench | — | 36.2% |
| GBAEval | — | 52.6% |
| LMArena Search | — | 1257 |
Reasoning GPT-5.6 Sol leads
GPT-5.6 Luna: 47.6 (#43), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-2 | 59.5% | 92.5% |
| SimpleBench | 46.8% | 71.7% |
| Kagi LLM Benchmark | 49.1% | 67% |
| NYT Connections (extended) | 69.4% | 93.8% |
| ARC-AGI-1 | 88% | 97.5% |
| CritPt | 20.6% | 32.3% |
| Chess Puzzles | 40% | 64% |
| LMArena Hard Prompts | 1451 | 1484 |
| Mystery Game Puzzles | 21% | 58% |
| DTBench | 89.1% | 96% |
| LMCA | 48.5% | 59.2% |
| Surface Evolver Bench | 61.9% | 93.1% |
| Epoch Capabilities Index | 156.39 | 161.66 |
| EnigmaEval | — | 37.1% |
| EBR-Bench | — | 44.8% |
| Bench to the Future 3 | — | 0.14 |
Math GPT-5.6 Sol leads
GPT-5.6 Luna: 77.7 (#14), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 82.1% | 89.1% |
| FrontierMath Tier 4 | 61% | 82.9% |
| OTIS Mock AIME 2024-2025 | 98.3% | 100% |
| ProofBench | 60% | 83% |
| LMArena Math | 1458 | 1474 |
| FrontierMath Erdős | — | 0% |
Knowledge GPT-5.6 Sol leads
GPT-5.6 Luna: 58.5 (#34), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 91.6% | 93.5% |
| SimpleQA Verified | 41% | 69.7% |
| LMArena Expert | 1478 | 1516 |
| Vectara Hallucination Rate | — | 12.4% |
Multimodal GPT-5.6 Sol leads
GPT-5.6 Luna: 42.7 (#28), GPT-5.6 Sol: 48.6 (#9)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | 1258 | 1281 |
| Blueprint-Bench 2 | 22.6% | 33.6% |
| Furniture Assembly | 42.5% | 56.7% |
| LMArena Document | 1457 | 1483 |
Multilingual GPT-5.6 Sol leads
GPT-5.6 Luna: 52.8 (#78), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1417 | 1452 |
| LMArena Chinese | 1470 | 1527 |
| LMArena French | 1456 | 1477 |
| LMArena German | 1454 | 1476 |
| LMArena Japanese | 1411 | 1471 |
| LMArena Korean | 1415 | 1442 |
| LMArena Russian | 1428 | 1468 |
| LMArena Spanish | 1448 | 1441 |
Instruction Following GPT-5.6 Sol leads
GPT-5.6 Luna: 75.6 (#57), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1437 | 1482 |
Long Context GPT-5.6 Sol leads
GPT-5.6 Luna: 43.9 (#82), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1436 | 1480 |
Writing & Preference GPT-5.6 Sol leads
GPT-5.6 Luna: 68.0 (#29), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | GPT-5.6 Luna | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1431 | 1457 |
| LMArena Creative Writing | 1396 | 1448 |
| EQ-Bench Creative Writing | 1829 | 1972 |
| EQ-Bench 4 | 1156 | 1250 |
| LMArena Multi-Turn | 1434 | 1460 |
Frequently asked questions
Is GPT-5.6 Luna better than GPT-5.6 Sol?
GPT-5.6 Sol is the stronger model overall, scoring 65.0 to 54.6 on the Noometry Index. GPT-5.6 Luna costs 18× 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 Luna or GPT-5.6 Sol?
GPT-5.6 Luna is cheaper. It lists at $0.20 per million input tokens and $1.20 per million output tokens; GPT-5.6 Sol lists at $4 and $20.
Is GPT-5.6 Luna or GPT-5.6 Sol better for coding?
GPT-5.6 Sol scores higher on coding benchmarks: 65.1 versus 54.5 in the Noometry coding category.
Which has the bigger context window?
Both accept 1.05M tokens.
How many benchmarks do GPT-5.6 Luna and GPT-5.6 Sol share?
52 benchmarks have published results for both models. GPT-5.6 Luna has 52 scored results on Noometry and GPT-5.6 Sol has 65.