Model comparison
GPT-5.4 mini vs o3-mini
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 36.7 on the Noometry Index.
Last verified . 35 shared benchmarks.
Summary
- They share 35 benchmarks with published results for both. GPT-5.4 mini scores higher in 8 categories and o3-mini in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 mini leads 45.5 to 28.1.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 51.2% for GPT-5.4 mini and 18.6% for o3-mini.
- GPT-5.4 mini is cheaper at $0.75 / $4.50 per million input/output tokens, against $1.10 / $4.40 for o3-mini.
- GPT-5.4 mini accepts more context: 400K tokens versus 200K.
Side by side
| GPT-5.4 mini | o3-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 45.0 | 36.7 |
| Released | 2026-03-17 | 2024-12-20 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $0.75 | $1.10 |
| Output $ / M tokens | $4.50 | $4.40 |
| Results tracked | 46 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GPT-5.4 mini: 45.2 (#72), o3-mini: 40.8 (#132)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| SciCode | 49.9% | 39.8% |
| WeirdML | 60.3% | 43.7% |
| LMArena Coding | 1438 | 1378 |
| FrontierCode | 27% | — |
| Aider Polyglot | — | 60.4% |
| LMArena WebDev | 1397 | — |
| GSO | — | 1.3% |
| LiveBench Coding | — | 82.7% |
| CadEval | — | 54% |
| ALE-Bench | 1,189 | — |
Agentic & Tool Use Too close to call
GPT-5.4 mini: 29.9 (#81), o3-mini: 29.6 (#84)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| Cybench | — | 22.5% |
| DeepResearch Bench | 36.3% | — |
Reasoning GPT-5.4 mini leads
GPT-5.4 mini: 30.4 (#85), o3-mini: 16.3 (#305)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| ARC-AGI-2 | 18.9% | 3% |
| ARC-AGI-1 | 63.7% | 34.5% |
| CritPt | 10% | 0.3% |
| Chess Puzzles | 24% | 17% |
| LMArena Hard Prompts | 1424 | 1366 |
| Mystery Game Puzzles | 11% | 7% |
| DTBench | 80% | 68.8% |
| LMCA | 40.8% | 19% |
| Epoch Capabilities Index | 148.84 | 140.34 |
| ForecastBench | 57 | 59.6 |
| SimpleBench | — | 22.8% |
| Kagi LLM Benchmark | 37.9% | — |
| NYT Connections (extended) | 61.8% | — |
| Thematic Generalization | 61.7% | — |
| LiveBench Reasoning | — | 89.6% |
| LiveBench Data Analysis | — | 70.6% |
| LiveBench | — | 75.9% |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.5 (#75), o3-mini: 28.1 (#244)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 51.2% | 18.6% |
| FrontierMath Tier 4 | 9.8% | 0% |
| OTIS Mock AIME 2024-2025 | 88.9% | 76.9% |
| LMArena Math | 1419 | 1396 |
| FrontierMath (Feb 2025 set) | 28.3% | 12.4% |
| FrontierMath Tier 4 (v1) | 2.1% | 4.2% |
| ProofBench | 21% | — |
| LiveBench Math | — | 77.3% |
| MATH Level 5 | — | 96.5% |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.5 (#67), o3-mini: 38.3 (#146)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| GPQA Diamond | 86.9% | 77% |
| SimpleQA Verified | 29.4% | 15.3% |
| LMArena Expert | 1435 | 1364 |
| Confabulations | — | 17.9% |
| Vectara Hallucination Rate | 5.5% | — |
Multimodal Not comparable
GPT-5.4 mini: 39.7 (#56), o3-mini: —
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual GPT-5.4 mini leads
GPT-5.4 mini: 51.9 (#96), o3-mini: 45.7 (#164)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| LMArena Non-English | 1405 | 1319 |
| LMArena Chinese | 1446 | 1379 |
| LMArena French | 1440 | 1334 |
| LMArena German | 1409 | 1303 |
| LMArena Japanese | 1374 | 1286 |
| LMArena Korean | 1368 | 1314 |
| LMArena Russian | 1417 | 1304 |
| LMArena Spanish | 1405 | 1321 |
Instruction Following Too close to call
GPT-5.4 mini: 74.1 (#102), o3-mini: 75.1 (#72)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| LMArena Instruction Following | 1405 | 1337 |
| LiveBench Instruction Following | — | 84.4% |
Long Context GPT-5.4 mini leads
GPT-5.4 mini: 43.0 (#112), o3-mini: 33.8 (#256)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| LMArena Longer Query | 1407 | 1343 |
| Fiction.LiveBench | — | 50% |
Writing & Preference GPT-5.4 mini leads
GPT-5.4 mini: 64.0 (#58), o3-mini: 50.3 (#182)
| Benchmark | GPT-5.4 mini | o3-mini |
|---|---|---|
| LMArena Text | 1412 | 1337 |
| LMArena Creative Writing | 1370 | 1286 |
| LMArena Multi-Turn | 1429 | 1320 |
| Short-Story Creative Writing | — | 61.7% |
| EQ-Bench Creative Writing | 1665 | — |
| LiveBench Language | — | 50.7% |
Frequently asked questions
Is GPT-5.4 mini better than o3-mini?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 36.7 on the Noometry Index.
Which is cheaper, GPT-5.4 mini or o3-mini?
GPT-5.4 mini is cheaper. It lists at $0.75 per million input tokens and $4.50 per million output tokens; o3-mini lists at $1.10 and $4.40.
Is GPT-5.4 mini or o3-mini better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 40.8 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 mini does, with 400K tokens against 200K.
How many benchmarks do GPT-5.4 mini and o3-mini share?
35 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and o3-mini has 51.