Model comparison
GPT-5.6 Luna vs GPT-6 Luna
GPT-5.6 Luna is the stronger model overall, scoring 54.6 to 53.3 on the Noometry Index. GPT-6 Luna costs 2.3× less per token, which makes it the better buy when GPT-5.6 Luna's lead doesn't matter for your workload.
Last verified . 42 shared benchmarks.
Summary
- They share 42 benchmarks with published results for both. GPT-5.6 Luna scores higher in 8 categories and GPT-6 Luna in 2 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-5.6 Luna leads 68.0 to 58.3.
- The biggest single-benchmark swing is Mystery Game Puzzles: 21% for GPT-5.6 Luna and 7% for GPT-6 Luna.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $0.20 / $1.20 for GPT-5.6 Luna.
Side by side
| GPT-5.6 Luna | GPT-6 Luna | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 54.6 | 53.3 |
| Released | 2026-07-09 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.20 | $0.10 |
| Output $ / M tokens | $1.20 | $0.50 |
| Results tracked | 52 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
GPT-5.6 Luna: 54.5 (#28), GPT-6 Luna: 55.5 (#25)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| DeepSWE | 67.2% | 66.6% |
| FrontierCode | 39.8% | 42.4% |
| LMArena WebDev | 1519 | 1581 |
| SciCode | 53.6% | 54.6% |
| LMArena Coding | 1466 | 1439 |
| ALE-Bench | 1,667 | 1,577 |
| CursorBench | 35.9% | — |
| WeirdML | 60.9% | — |
Agentic & Tool Use GPT-5.6 Luna leads
GPT-5.6 Luna: 34.4 (#45), GPT-6 Luna: 33.3 (#54)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| APEX-Agents | 43% | 44.3% |
| GDP.pdf | 22.7% | 23% |
| BALROG | 45.6% | — |
| Vending-Bench 2 | 4,095 | — |
Reasoning Too close to call
GPT-5.6 Luna: 47.6 (#43), GPT-6 Luna: 48.2 (#41)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| ARC-AGI-2 | 59.5% | 59.3% |
| NYT Connections (extended) | 69.4% | 68.7% |
| ARC-AGI-1 | 88% | 86.7% |
| CritPt | 20.6% | 19.4% |
| Chess Puzzles | 40% | 31% |
| LMArena Hard Prompts | 1451 | 1411 |
| Mystery Game Puzzles | 21% | 7% |
| DTBench | 89.1% | 90.1% |
| LMCA | 48.5% | 44.5% |
| Epoch Capabilities Index | 156.39 | 156.28 |
| SimpleBench | 46.8% | — |
| Kagi LLM Benchmark | 49.1% | — |
| Surface Evolver Bench | 61.9% | — |
Math GPT-5.6 Luna leads
GPT-5.6 Luna: 77.7 (#14), GPT-6 Luna: 76.1 (#15)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| FrontierMath (Tiers 1-3) | 82.1% | 78.9% |
| FrontierMath Tier 4 | 61% | 56.1% |
| OTIS Mock AIME 2024-2025 | 98.3% | 98.9% |
| ProofBench | 60% | 64% |
| LMArena Math | 1458 | 1416 |
Knowledge GPT-5.6 Luna leads
GPT-5.6 Luna: 58.5 (#34), GPT-6 Luna: 57.0 (#41)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 91.6% | 90.5% |
| SimpleQA Verified | 41% | 41.4% |
| LMArena Expert | 1478 | 1444 |
Multimodal Too close to call
GPT-5.6 Luna: 42.7 (#28), GPT-6 Luna: 42.4 (#30)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| LMArena Vision | 1258 | 1217 |
| Blueprint-Bench 2 | 22.6% | 31.2% |
| Furniture Assembly | 42.5% | 44.2% |
| LMArena Document | 1457 | — |
Multilingual GPT-5.6 Luna leads
GPT-5.6 Luna: 52.8 (#78), GPT-6 Luna: 50.5 (#117)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1417 | 1386 |
| LMArena Chinese | 1470 | 1433 |
| LMArena French | 1456 | 1420 |
| LMArena German | 1454 | 1369 |
| LMArena Japanese | 1411 | 1369 |
| LMArena Korean | 1415 | 1360 |
| LMArena Russian | 1428 | 1394 |
| LMArena Spanish | 1448 | 1393 |
Instruction Following GPT-5.6 Luna leads
GPT-5.6 Luna: 75.6 (#57), GPT-6 Luna: 74.3 (#99)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1437 | 1409 |
Long Context Too close to call
GPT-5.6 Luna: 43.9 (#82), GPT-6 Luna: 43.0 (#111)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1436 | 1409 |
Writing & Preference GPT-5.6 Luna leads
GPT-5.6 Luna: 68.0 (#29), GPT-6 Luna: 58.3 (#119)
| Benchmark | GPT-5.6 Luna | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1431 | 1391 |
| LMArena Creative Writing | 1396 | 1363 |
| LMArena Multi-Turn | 1434 | 1396 |
| EQ-Bench Creative Writing | 1829 | — |
| EQ-Bench 4 | 1156 | — |
Frequently asked questions
Is GPT-5.6 Luna better than GPT-6 Luna?
GPT-5.6 Luna is the stronger model overall, scoring 54.6 to 53.3 on the Noometry Index. GPT-6 Luna costs 2.3× less per token, which makes it the better buy when GPT-5.6 Luna's lead doesn't matter for your workload.
Which is cheaper, GPT-5.6 Luna or GPT-6 Luna?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; GPT-5.6 Luna lists at $0.20 and $1.20.
Is GPT-5.6 Luna or GPT-6 Luna better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 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-6 Luna share?
42 benchmarks have published results for both models. GPT-5.6 Luna has 52 scored results on Noometry and GPT-6 Luna has 42.