Model comparison
GPT-5.2 vs GPT-6 Luna
GPT-5.2 and GPT-6 Luna score almost the same on the Noometry Index (54.1 vs 53.3), so choose on price, context window or the category you care about most.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. GPT-5.2 scores higher in 8 categories and GPT-6 Luna in 2 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Luna leads 76.1 to 60.0.
- The biggest single-benchmark swing is ProofBench: 15% for GPT-5.2 and 64% for GPT-6 Luna.
- GPT-6 Luna is cheaper at $0.10 / $0.50 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
- GPT-6 Luna accepts more context: 1.05M tokens versus 400K.
Side by side
| GPT-5.2 | GPT-6 Luna | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 54.1 | 53.3 |
| Released | 2025-12-11 | 2026-09-22 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $1.75 | $0.10 |
| Output $ / M tokens | $14 | $0.50 |
| Results tracked | 67 | 42 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Luna leads
GPT-5.2: 51.6 (#37), GPT-6 Luna: 55.5 (#25)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena WebDev | 1416 | 1581 |
| LMArena Coding | 1447 | 1439 |
| ALE-Bench | 1,294 | 1,577 |
| SWE-bench Verified | 73.8% | — |
| DeepSWE | — | 66.6% |
| FrontierCode | — | 42.4% |
| SWE-bench Verified (bash only) | 72.8% | — |
| SWE-bench Multilingual | 66.7% | — |
| SciCode | — | 54.6% |
| GSO | 27.4% | — |
| WeirdML | 72.2% | — |
| AlgoTune | 2.05 | — |
Agentic & Tool Use GPT-5.2 leads
GPT-5.2: 40.2 (#24), GPT-6 Luna: 33.3 (#54)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| Terminal-Bench | 64.9% | — |
| APEX-Agents | — | 44.3% |
| Berkeley Function Calling Leaderboard | 55.9% | — |
| GDPval | 49.7% | — |
| Remote Labor Index | 2.5% | — |
| τ²-bench Airline | 83% | — |
| τ²-bench Banking | 32.2% | — |
| τ²-bench Retail | 81.6% | — |
| τ²-bench Telecom | 89.7% | — |
| DeepResearch Bench | 41.1% | — |
| GDP.pdf | — | 23% |
| LMArena Search | 1207 | — |
| METR Time Horizons | 75.3% | — |
| Vending-Bench 2 | 3,591 | — |
Reasoning GPT-5.2 leads
GPT-5.2: 50.2 (#35), GPT-6 Luna: 48.2 (#41)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| ARC-AGI-2 | 52.9% | 59.3% |
| NYT Connections (extended) | 83.6% | 68.7% |
| ARC-AGI-1 | 86.2% | 86.7% |
| Chess Puzzles | 49% | 31% |
| LMArena Hard Prompts | 1445 | 1411 |
| Mystery Game Puzzles | 23% | 7% |
| DTBench | 90.9% | 90.1% |
| LMCA | 43.9% | 44.5% |
| Epoch Capabilities Index | 153.45 | 156.28 |
| SimpleBench | 45.8% | — |
| Kagi LLM Benchmark | 73.3% | — |
| CritPt | — | 19.4% |
| EnigmaEval | 10.4% | — |
| EBR-Bench | 23% | — |
| ForecastBench | 60.1 | — |
Math GPT-6 Luna leads
GPT-5.2: 60.0 (#38), GPT-6 Luna: 76.1 (#15)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 78.9% |
| FrontierMath Tier 4 | 31.7% | 56.1% |
| OTIS Mock AIME 2024-2025 | 96.1% | 98.9% |
| ProofBench | 15% | 64% |
| LMArena Math | 1440 | 1416 |
| MathArena Final-Answer Competitions | 72% | — |
| FrontierMath (Feb 2025 set) | 40.7% | — |
| FrontierMath Tier 4 (v1) | 18.8% | — |
Knowledge GPT-5.2 leads
GPT-5.2: 59.3 (#32), GPT-6 Luna: 57.0 (#41)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| GPQA Diamond | 91.4% | 90.5% |
| SimpleQA Verified | 37.1% | 41.4% |
| LMArena Expert | 1445 | 1444 |
| Humanity's Last Exam | 27.8% | — |
| Vectara Hallucination Rate | 8.4% | — |
Multimodal GPT-5.2 leads
GPT-5.2: 51.3 (#7), GPT-6 Luna: 42.4 (#30)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena Vision | 1268 | 1217 |
| Furniture Assembly | 38.3% | 44.2% |
| VPCT | 84% | — |
| Blueprint-Bench 2 | — | 31.2% |
| LMArena Document | 1405 | — |
Multilingual GPT-5.2 leads
GPT-5.2: 53.4 (#67), GPT-6 Luna: 50.5 (#117)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena Non-English | 1425 | 1386 |
| LMArena Chinese | 1460 | 1433 |
| LMArena French | 1455 | 1420 |
| LMArena German | 1448 | 1369 |
| LMArena Japanese | 1420 | 1369 |
| LMArena Korean | 1392 | 1360 |
| LMArena Russian | 1440 | 1394 |
| LMArena Spanish | 1433 | 1393 |
Instruction Following Too close to call
GPT-5.2: 74.7 (#89), GPT-6 Luna: 74.3 (#99)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena Instruction Following | 1417 | 1409 |
Long Context GPT-5.2 leads
GPT-5.2: 44.0 (#78), GPT-6 Luna: 43.0 (#111)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena Longer Query | 1428 | 1409 |
| CL-bench | 18.2% | — |
Writing & Preference GPT-5.2 leads
GPT-5.2: 66.8 (#32), GPT-6 Luna: 58.3 (#119)
| Benchmark | GPT-5.2 | GPT-6 Luna |
|---|---|---|
| LMArena Text | 1439 | 1391 |
| LMArena Creative Writing | 1401 | 1363 |
| LMArena Multi-Turn | 1458 | 1396 |
| EQ-Bench Creative Writing | 1703 | — |
Frequently asked questions
Is GPT-5.2 better than GPT-6 Luna?
GPT-5.2 and GPT-6 Luna score almost the same on the Noometry Index (54.1 vs 53.3), so choose on price, context window or the category you care about most.
Which is cheaper, GPT-5.2 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-5.2 lists at $1.75 and $14.
Is GPT-5.2 or GPT-6 Luna better for coding?
GPT-6 Luna scores higher on coding benchmarks: 55.5 versus 51.6 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Luna does, with 1.05M tokens against 400K.
How many benchmarks do GPT-5.2 and GPT-6 Luna share?
35 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and GPT-6 Luna has 42.