Model comparison
GPT-6 Luna vs o4-mini
GPT-6 Luna is the stronger model overall, scoring 53.3 to 41.6 on the Noometry Index.
Last verified . 32 shared benchmarks.
Summary
- They share 32 benchmarks with published results for both. GPT-6 Luna scores higher in 8 categories and o4-mini in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 40.8.
- The biggest single-benchmark swing is ARC-AGI-2: 59.3% for GPT-6 Luna and 6.1% for o4-mini.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $1.10 / $4.40 for o4-mini.
- GPT-6 Luna accepts more context: 1.05M tokens versus 200K.
Side by side
| GPT-6 Luna | o4-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 53.3 | 41.6 |
| Released | 2026-09-22 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $0.10 | $1.10 |
| Output $ / M tokens | $0.50 | $4.40 |
| Results tracked | 42 | 60 |
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), o4-mini: 40.9 (#127)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Coding | 1439 | 1368 |
| ALE-Bench | 1,577 | 826.17 |
| DeepSWE | 66.6% | — |
| FrontierCode | 42.4% | — |
| SWE-bench Verified (bash only) | — | 45% |
| Aider Polyglot | — | 72% |
| LMArena WebDev | 1581 | — |
| SciCode | 54.6% | — |
| GSO | — | 3.6% |
| WeirdML | — | 52.6% |
| CadEval | — | 62% |
| AlgoTune | — | 1.72 |
Agentic & Tool Use Too close to call
GPT-6 Luna: 33.3 (#54), o4-mini: 32.6 (#61)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| APEX-Agents | 44.3% | — |
| Berkeley Function Calling Leaderboard | — | 53.2% |
| GDPval | — | 25.3% |
| GDP.pdf | 23% | — |
| METR Time Horizons | — | 63.9% |
Reasoning GPT-6 Luna leads
GPT-6 Luna: 48.2 (#41), o4-mini: 24.6 (#162)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| ARC-AGI-2 | 59.3% | 6.1% |
| ARC-AGI-1 | 86.7% | 58.7% |
| CritPt | 19.4% | 0.6% |
| Chess Puzzles | 31% | 26% |
| LMArena Hard Prompts | 1411 | 1351 |
| Mystery Game Puzzles | 7% | 5% |
| DTBench | 90.1% | 77.6% |
| LMCA | 44.5% | 26.5% |
| Epoch Capabilities Index | 156.28 | 145.64 |
| SimpleBench | — | 38.7% |
| Kagi LLM Benchmark | — | 67.6% |
| NYT Connections (extended) | 68.7% | — |
| EnigmaEval | — | 9.2% |
| ForecastBench | — | 61.8 |
Math GPT-6 Luna leads
GPT-6 Luna: 76.1 (#15), o4-mini: 40.8 (#89)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 78.9% | 36.1% |
| FrontierMath Tier 4 | 56.1% | 4.9% |
| OTIS Mock AIME 2024-2025 | 98.9% | 81.7% |
| LMArena Math | 1416 | 1389 |
| ProofBench | 64% | — |
| Omni-MATH | — | 72% |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 24.8% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge GPT-6 Luna leads
GPT-6 Luna: 57.0 (#41), o4-mini: 43.6 (#91)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| GPQA Diamond | 90.5% | 79.6% |
| SimpleQA Verified | 41.4% | 19.6% |
| LMArena Expert | 1444 | 1343 |
| Humanity's Last Exam | — | 18.1% |
| MMLU-Pro | — | 82% |
| Confabulations | — | 15.8% |
| Vectara Hallucination Rate | — | 18.6% |
| GPQA (HELM) | — | 73.5% |
Multimodal GPT-6 Luna leads
GPT-6 Luna: 42.4 (#30), o4-mini: 40.2 (#49)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Vision | 1217 | 1194 |
| GeoBench | — | 64% |
| VPCT | — | 57.5% |
| Blueprint-Bench 2 | 31.2% | — |
| Furniture Assembly | 44.2% | — |
Multilingual GPT-6 Luna leads
GPT-6 Luna: 50.5 (#117), o4-mini: 47.0 (#154)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Non-English | 1386 | 1337 |
| LMArena Chinese | 1433 | 1354 |
| LMArena French | 1420 | 1364 |
| LMArena German | 1369 | 1336 |
| LMArena Japanese | 1369 | 1308 |
| LMArena Korean | 1360 | 1312 |
| LMArena Russian | 1394 | 1334 |
| LMArena Spanish | 1393 | 1347 |
Instruction Following Too close to call
GPT-6 Luna: 74.3 (#99), o4-mini: 75.2 (#68)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Instruction Following | 1409 | 1321 |
| IFEval | — | 92.8% |
Long Context o4-mini leads
GPT-6 Luna: 43.0 (#111), o4-mini: 45.5 (#33)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Longer Query | 1409 | 1315 |
| Fiction.LiveBench | — | 77.8% |
Writing & Preference GPT-6 Luna leads
GPT-6 Luna: 58.3 (#119), o4-mini: 54.0 (#152)
| Benchmark | GPT-6 Luna | o4-mini |
|---|---|---|
| LMArena Text | 1391 | 1353 |
| LMArena Creative Writing | 1363 | 1294 |
| LMArena Multi-Turn | 1396 | 1350 |
| Short-Story Creative Writing | — | 75% |
| WildBench | — | 85.4% |
Frequently asked questions
Is GPT-6 Luna better than o4-mini?
GPT-6 Luna is the stronger model overall, scoring 53.3 to 41.6 on the Noometry Index.
Which is cheaper, GPT-6 Luna or o4-mini?
GPT-6 Luna is cheaper. It lists at $0.10 per million input tokens and $0.50 per million output tokens; o4-mini lists at $1.10 and $4.40.
Is GPT-6 Luna or o4-mini better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 40.9 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 o4-mini share?
32 benchmarks have published results for both models. GPT-6 Luna has 42 scored results on Noometry and o4-mini has 60.