Model comparison
GPT-4.5 vs GPT-5 Mini
GPT-5 Mini is the stronger model overall, scoring 41.8 to 37.2 on the Noometry Index.
Last verified . 32 shared benchmarks.
Summary
- They share 32 benchmarks with published results for both. GPT-4.5 scores higher in 4 categories and GPT-5 Mini in 6 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5 Mini leads 46.7 to 32.6.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 37.8% for GPT-4.5 and 86.7% for GPT-5 Mini.
Side by side
| GPT-4.5 | GPT-5 Mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 37.2 | 41.8 |
| Released | 2025-02-27 | 2025-08-07 |
| Weights | Proprietary | Proprietary |
| Context window | — | 400K |
| Max output | — | 128K |
| Input $ / M tokens | — | $0.25 |
| Output $ / M tokens | — | $2 |
| Results tracked | 42 | 60 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-4.5 leads
GPT-4.5: 42.2 (#109), GPT-5 Mini: 40.1 (#146)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| WeirdML | 39.4% | 52.7% |
| LMArena Coding | 1396 | 1406 |
| SWE-bench Verified | — | 64.7% |
| SWE-bench Verified (bash only) | — | 59.8% |
| Aider Polyglot | 44.9% | — |
| SWE-bench Multilingual | — | 39.7% |
| SciCode | — | 39.2% |
| LiveBench Coding | 75.2% | — |
| ALE-Bench | — | 799.77 |
| AlgoTune | — | 1.38 |
Agentic & Tool Use GPT-5 Mini leads
GPT-4.5: 27.9 (#97), GPT-5 Mini: 31.1 (#70)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| Terminal-Bench | — | 34.8% |
| Berkeley Function Calling Leaderboard | — | 55.5% |
| Cybench | 17.5% | — |
| Vending-Bench 2 | — | -31.18 |
Reasoning GPT-5 Mini leads
GPT-4.5: 13.9 (#330), GPT-5 Mini: 23.9 (#168)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| ARC-AGI-2 | 0.8% | 4.4% |
| ARC-AGI-1 | 10.3% | 54.3% |
| EnigmaEval | 3.2% | 8.2% |
| LMArena Hard Prompts | 1403 | 1380 |
| Epoch Capabilities Index | 136.74 | 145.52 |
| ForecastBench | 61.7 | 61 |
| SimpleBench | 34.5% | — |
| Kagi LLM Benchmark | — | 70.3% |
| CritPt | — | 0% |
| Chess Puzzles | — | 30% |
| LiveBench Reasoning | 71.1% | — |
| Mystery Game Puzzles | — | 10% |
| DTBench | — | 80.5% |
| LiveBench Data Analysis | 64.3% | — |
| LMCA | — | 34.2% |
| LiveBench | 69% | — |
Math GPT-5 Mini leads
GPT-4.5: 32.6 (#211), GPT-5 Mini: 46.7 (#69)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 37.8% | 86.7% |
| LMArena Math | 1412 | 1378 |
| MATH Level 5 | 78.6% | 97.8% |
| FrontierMath (Tiers 1-3) | — | 46.7% |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 9% |
| Omni-MATH | — | 72.2% |
| LiveBench Math | 69.3% | — |
| FrontierMath (Feb 2025 set) | — | 27.2% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge GPT-5 Mini leads
GPT-4.5: 32.5 (#211), GPT-5 Mini: 45.6 (#86)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| GPQA Diamond | 68.7% | 75% |
| Humanity's Last Exam | 5.4% | 19.4% |
| Confabulations | 13.6% | 13.3% |
| LMArena Expert | 1394 | 1379 |
| SimpleQA Verified | — | 21.6% |
| MMLU-Pro | — | 83.5% |
| Vectara Hallucination Rate | — | 12.9% |
| GPQA (HELM) | — | 75.6% |
Multimodal GPT-4.5 leads
GPT-4.5: 37.6 (#71), GPT-5 Mini: 35.6 (#85)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Vision | 1195 | 1202 |
| VPCT | 45% | 40.2% |
Multilingual GPT-4.5 leads
GPT-4.5: 52.5 (#83), GPT-5 Mini: 48.9 (#137)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Non-English | 1413 | 1363 |
| LMArena Chinese | 1421 | 1385 |
| LMArena French | 1418 | 1386 |
| LMArena German | 1457 | 1366 |
| LMArena Japanese | 1416 | 1341 |
| LMArena Korean | 1392 | 1308 |
| LMArena Russian | 1419 | 1362 |
| LMArena Spanish | — | 1355 |
Instruction Following GPT-5 Mini leads
GPT-4.5: 72.6 (#134), GPT-5 Mini: 76.2 (#46)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Instruction Following | 1404 | 1357 |
| LiveBench Instruction Following | 72.3% | — |
| IFEval | — | 92.7% |
Long Context GPT-5 Mini leads
GPT-4.5: 40.4 (#155), GPT-5 Mini: 41.9 (#132)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| Fiction.LiveBench | 63.9% | 69.4% |
| LMArena Longer Query | 1406 | 1355 |
Writing & Preference GPT-4.5 leads
GPT-4.5: 56.9 (#134), GPT-5 Mini: 55.2 (#148)
| Benchmark | GPT-4.5 | GPT-5 Mini |
|---|---|---|
| LMArena Text | 1417 | 1373 |
| LMArena Creative Writing | 1394 | 1325 |
| Short-Story Creative Writing | 75.6% | 83.1% |
| EQ-Bench Creative Writing | 1258 | 1313 |
| LMArena Multi-Turn | 1444 | 1363 |
| WildBench | — | 85.5% |
| LiveBench Language | 61.5% | — |
Frequently asked questions
Is GPT-4.5 better than GPT-5 Mini?
GPT-5 Mini is the stronger model overall, scoring 41.8 to 37.2 on the Noometry Index.
Is GPT-4.5 or GPT-5 Mini better for coding?
GPT-4.5 scores higher on coding benchmarks: 42.2 versus 40.1 in the Noometry coding category.
How many benchmarks do GPT-4.5 and GPT-5 Mini share?
32 benchmarks have published results for both models. GPT-4.5 has 42 scored results on Noometry and GPT-5 Mini has 60.