Model comparison
GPT-6 Luna vs o3
GPT-6 Luna is the stronger model overall, scoring 53.3 to 47.5 on the Noometry Index.
Last verified . 31 shared benchmarks.
Summary
- They share 31 benchmarks with published results for both. GPT-6 Luna scores higher in 6 categories and o3 in 4 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 50.2.
- The biggest single-benchmark swing is ARC-AGI-2: 59.3% for GPT-6 Luna and 6.5% for o3.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $2 / $8 for o3.
- GPT-6 Luna accepts more context: 1.05M tokens versus 200K.
Side by side
| GPT-6 Luna | o3 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 53.3 | 47.5 |
| Released | 2026-09-22 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $0.10 | $2 |
| Output $ / M tokens | $0.50 | $8 |
| Results tracked | 42 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
GPT-6 Luna: 55.5 (#25), o3: 46.8 (#64)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Coding | 1439 | 1408 |
| ALE-Bench | 1,577 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| DeepSWE | 66.6% | — |
| FrontierCode | 42.4% | — |
| SWE-bench Verified (bash only) | — | 58.4% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1581 | — |
| SciCode | 54.6% | — |
| GSO | — | 8.8% |
| WeirdML | — | 52.4% |
| CadEval | — | 74% |
Agentic & Tool Use o3 leads
GPT-6 Luna: 33.3 (#54), o3: 34.5 (#44)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| APEX-Agents | 44.3% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| GDP.pdf | 23% | — |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
Reasoning GPT-6 Luna leads
GPT-6 Luna: 48.2 (#41), o3: 32.0 (#78)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| ARC-AGI-2 | 59.3% | 6.5% |
| ARC-AGI-1 | 86.7% | 60.8% |
| CritPt | 19.4% | 1.4% |
| Chess Puzzles | 31% | 38% |
| LMArena Hard Prompts | 1411 | 1402 |
| Mystery Game Puzzles | 7% | 29% |
| DTBench | 90.1% | 84.8% |
| LMCA | 44.5% | 39.7% |
| Epoch Capabilities Index | 156.28 | 146.86 |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 68.7% | — |
| EnigmaEval | — | 13.1% |
| ForecastBench | — | 62.5 |
Math GPT-6 Luna leads
GPT-6 Luna: 76.1 (#15), o3: 50.2 (#58)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 78.9% | 33.3% |
| OTIS Mock AIME 2024-2025 | 98.9% | 84.4% |
| LMArena Math | 1416 | 1426 |
| FrontierMath Tier 4 | 56.1% | — |
| ProofBench | 64% | — |
| Omni-MATH | — | 71.4% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge GPT-6 Luna leads
GPT-6 Luna: 57.0 (#41), o3: 54.6 (#52)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| GPQA Diamond | 90.5% | 81.8% |
| SimpleQA Verified | 41.4% | 49.4% |
| LMArena Expert | 1444 | 1402 |
| Humanity's Last Exam | — | 20.3% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
Multimodal GPT-6 Luna leads
GPT-6 Luna: 42.4 (#30), o3: 41.4 (#36)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Vision | 1217 | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
| Blueprint-Bench 2 | 31.2% | — |
| Furniture Assembly | 44.2% | — |
Multilingual o3 leads
GPT-6 Luna: 50.5 (#117), o3: 51.7 (#105)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Non-English | 1386 | 1401 |
| LMArena Chinese | 1433 | 1437 |
| LMArena French | 1420 | 1430 |
| LMArena German | 1369 | 1420 |
| LMArena Japanese | 1369 | 1403 |
| LMArena Korean | 1360 | 1370 |
| LMArena Russian | 1394 | 1406 |
| LMArena Spanish | 1393 | 1395 |
Instruction Following GPT-6 Luna leads
GPT-6 Luna: 74.3 (#99), o3: 72.8 (#127)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Instruction Following | 1409 | 1368 |
| IFEval | — | 86.9% |
Long Context o3 leads
GPT-6 Luna: 43.0 (#111), o3: 53.3 (#6)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Longer Query | 1409 | 1372 |
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
Writing & Preference o3 leads
GPT-6 Luna: 58.3 (#119), o3: 63.5 (#64)
| Benchmark | GPT-6 Luna | o3 |
|---|---|---|
| LMArena Text | 1391 | 1410 |
| LMArena Creative Writing | 1363 | 1359 |
| LMArena Multi-Turn | 1396 | 1405 |
| Short-Story Creative Writing | — | 83.9% |
| EQ-Bench Creative Writing | — | 1676 |
| WildBench | — | 86.1% |
Frequently asked questions
Is GPT-6 Luna better than o3?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 47.5 on the Noometry Index.
Which is cheaper, GPT-6 Luna or o3?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; o3 lists at $2 and $8.
Is GPT-6 Luna or o3 better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 46.8 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 200K.
How many benchmarks do GPT-6 Luna and o3 share?
31 benchmarks have published results for both models. GPT-6 Luna has 42 scored results on Noometry and o3 has 63.