Model comparison
GPT-5.2 vs GPT-5 Mini
GPT-5.2 is the stronger model overall, scoring 54.1 to 41.8 on the Noometry Index. GPT-5 Mini costs 7.0× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Last verified . 49 shared benchmarks.
Summary
- They share 49 benchmarks with published results for both. GPT-5.2 scores higher in 9 categories and GPT-5 Mini in 1 category; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-5.2 leads 50.2 to 23.9.
- The biggest single-benchmark swing is ARC-AGI-2: 52.9% for GPT-5.2 and 4.4% for GPT-5 Mini.
- GPT-5 Mini is cheaper at $0.25 / $2 per million input/output tokens, against $1.75 / $14 for GPT-5.2.
Side by side
| GPT-5.2 | GPT-5 Mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 54.1 | 41.8 |
| Released | 2025-12-11 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 400K |
| Max output | 128K | 128K |
| Input $ / M tokens | $1.75 | $0.25 |
| Output $ / M tokens | $14 | $2 |
| Results tracked | 67 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.2 leads
GPT-5.2: 51.6 (#37), GPT-5 Mini: 40.1 (#146)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| SWE-bench Verified | 73.8% | 64.7% |
| SWE-bench Verified (bash only) | 72.8% | 59.8% |
| SWE-bench Multilingual | 66.7% | 39.7% |
| WeirdML | 72.2% | 52.7% |
| LMArena Coding | 1447 | 1406 |
| ALE-Bench | 1,294 | 799.77 |
| AlgoTune | 2.05 | 1.38 |
| LMArena WebDev | 1416 | — |
| SciCode | — | 39.2% |
| GSO | 27.4% | — |
Agentic & Tool Use GPT-5.2 leads
GPT-5.2: 40.2 (#24), GPT-5 Mini: 31.1 (#70)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | 64.9% | 34.8% |
| Berkeley Function Calling Leaderboard | 55.9% | 55.5% |
| Vending-Bench 2 | 3,591 | -31.18 |
| 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% | — |
| LMArena Search | 1207 | — |
| METR Time Horizons | 75.3% | — |
Reasoning GPT-5.2 leads
GPT-5.2: 50.2 (#35), GPT-5 Mini: 23.9 (#168)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 52.9% | 4.4% |
| Kagi LLM Benchmark | 73.3% | 70.3% |
| ARC-AGI-1 | 86.2% | 54.3% |
| Chess Puzzles | 49% | 30% |
| EnigmaEval | 10.4% | 8.2% |
| LMArena Hard Prompts | 1445 | 1380 |
| Mystery Game Puzzles | 23% | 10% |
| DTBench | 90.9% | 80.5% |
| LMCA | 43.9% | 34.2% |
| Epoch Capabilities Index | 153.45 | 145.52 |
| ForecastBench | 60.1 | 61 |
| SimpleBench | 45.8% | — |
| NYT Connections (extended) | 83.6% | — |
| CritPt | — | 0% |
| EBR-Bench | 23% | — |
Math GPT-5.2 leads
GPT-5.2: 60.0 (#38), GPT-5 Mini: 46.7 (#69)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 46.7% |
| FrontierMath Tier 4 | 31.7% | 12.2% |
| OTIS Mock AIME 2024-2025 | 96.1% | 86.7% |
| ProofBench | 15% | 9% |
| LMArena Math | 1440 | 1378 |
| FrontierMath (Feb 2025 set) | 40.7% | 27.2% |
| FrontierMath Tier 4 (v1) | 18.8% | 6.3% |
| MathArena Final-Answer Competitions | 72% | — |
| Omni-MATH | — | 72.2% |
| MATH Level 5 | — | 97.8% |
Knowledge GPT-5.2 leads
GPT-5.2: 59.3 (#32), GPT-5 Mini: 45.6 (#86)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 91.4% | 75% |
| Humanity's Last Exam | 27.8% | 19.4% |
| SimpleQA Verified | 37.1% | 21.6% |
| Vectara Hallucination Rate | 8.4% | 12.9% |
| LMArena Expert | 1445 | 1379 |
| MMLU-Pro | — | 83.5% |
| Confabulations | — | 13.3% |
| GPQA (HELM) | — | 75.6% |
Multimodal GPT-5.2 leads
GPT-5.2: 51.3 (#7), GPT-5 Mini: 35.6 (#85)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1268 | 1202 |
| VPCT | 84% | 40.2% |
| Furniture Assembly | 38.3% | — |
| LMArena Document | 1405 | — |
Multilingual GPT-5.2 leads
GPT-5.2: 53.4 (#67), GPT-5 Mini: 48.9 (#137)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1425 | 1363 |
| LMArena Chinese | 1460 | 1385 |
| LMArena French | 1455 | 1386 |
| LMArena German | 1448 | 1366 |
| LMArena Japanese | 1420 | 1341 |
| LMArena Korean | 1392 | 1308 |
| LMArena Russian | 1440 | 1362 |
| LMArena Spanish | 1433 | 1355 |
Instruction Following GPT-5 Mini leads
GPT-5.2: 74.7 (#89), GPT-5 Mini: 76.2 (#46)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1417 | 1357 |
| IFEval | — | 92.7% |
Long Context GPT-5.2 leads
GPT-5.2: 44.0 (#78), GPT-5 Mini: 41.9 (#132)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| LMArena Longer Query | 1428 | 1355 |
| Fiction.LiveBench | — | 69.4% |
| CL-bench | 18.2% | — |
Writing & Preference GPT-5.2 leads
GPT-5.2: 66.8 (#32), GPT-5 Mini: 55.2 (#148)
| Benchmark | GPT-5.2 | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1439 | 1373 |
| LMArena Creative Writing | 1401 | 1325 |
| EQ-Bench Creative Writing | 1703 | 1313 |
| LMArena Multi-Turn | 1458 | 1363 |
| Short-Story Creative Writing | — | 83.1% |
| WildBench | — | 85.5% |
Frequently asked questions
Is GPT-5.2 better than GPT-5 Mini?
GPT-5.2 is the stronger model overall, scoring 54.1 to 41.8 on the Noometry Index. GPT-5 Mini costs 7.0× less per token, which makes it the better buy when GPT-5.2's lead doesn't matter for your workload.
Which is cheaper, GPT-5.2 or GPT-5 Mini?
GPT-5 Mini is cheaper. It lists at $0.25 per million input tokens and $2 per million output tokens; GPT-5.2 lists at $1.75 and $14.
Is GPT-5.2 or GPT-5 Mini better for coding?
GPT-5.2 scores higher on coding benchmarks: 51.6 versus 40.1 in the Noometry coding category.
Which has the bigger context window?
Both accept 400K tokens.
How many benchmarks do GPT-5.2 and GPT-5 Mini share?
49 benchmarks have published results for both models. GPT-5.2 has 67 scored results on Noometry and GPT-5 Mini has 60.