Model comparison
GPT-4.1 vs GPT-5.4 mini
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 35.9 on the Noometry Index.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. GPT-4.1 scores higher in 1 category and GPT-5.4 mini in 9 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 mini leads 45.5 to 22.3.
- The biggest single-benchmark swing is ARC-AGI-1: 5.5% for GPT-4.1 and 63.7% for GPT-5.4 mini.
- GPT-5.4 mini is cheaper at $0.75 / $4.50 per million input/output tokens, against $2 / $8 for GPT-4.1.
- GPT-4.1 accepts more context: 1.05M tokens versus 400K.
Side by side
| GPT-4.1 | GPT-5.4 mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 35.9 | 45.0 |
| Released | 2025-04-14 | 2026-03-17 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 400K |
| Max output | 33K | 128K |
| Input $ / M tokens | $2 | $0.75 |
| Output $ / M tokens | $8 | $4.50 |
| Results tracked | 52 | 46 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 mini leads
GPT-4.1: 34.4 (#238), GPT-5.4 mini: 45.2 (#72)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| WeirdML | 39% | 60.3% |
| LMArena Coding | 1391 | 1438 |
| ALE-Bench | 558.1 | 1,189 |
| SWE-bench Verified | 48.5% | — |
| FrontierCode | — | 27% |
| SWE-bench Verified (bash only) | 39.6% | — |
| Aider Polyglot | 52.4% | — |
| LMArena WebDev | — | 1397 |
| SciCode | — | 49.9% |
| CadEval | 42% | — |
Agentic & Tool Use GPT-4.1 leads
GPT-4.1: 34.7 (#43), GPT-5.4 mini: 29.9 (#81)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | 54% | — |
| DeepResearch Bench | — | 36.3% |
Reasoning GPT-5.4 mini leads
GPT-4.1: 11.7 (#339), GPT-5.4 mini: 30.4 (#85)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| ARC-AGI-2 | 0.4% | 18.9% |
| Kagi LLM Benchmark | 52.3% | 37.9% |
| ARC-AGI-1 | 5.5% | 63.7% |
| Chess Puzzles | 6% | 24% |
| LMArena Hard Prompts | 1384 | 1424 |
| DTBench | 68.3% | 80% |
| LMCA | 25.6% | 40.8% |
| Epoch Capabilities Index | 136.78 | 148.84 |
| ForecastBench | 61.5 | 57 |
| SimpleBench | 27% | — |
| NYT Connections (extended) | — | 61.8% |
| CritPt | — | 10% |
| EnigmaEval | 2.2% | — |
| Thematic Generalization | — | 61.7% |
| Mystery Game Puzzles | — | 11% |
Math GPT-5.4 mini leads
GPT-4.1: 22.3 (#280), GPT-5.4 mini: 45.5 (#75)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 6% | 51.2% |
| OTIS Mock AIME 2024-2025 | 38.3% | 88.9% |
| LMArena Math | 1370 | 1419 |
| FrontierMath (Feb 2025 set) | 5.5% | 28.3% |
| FrontierMath Tier 4 (v1) | 0% | 2.1% |
| FrontierMath Tier 4 | — | 9.8% |
| ProofBench | — | 21% |
| Omni-MATH | 47.1% | — |
| MATH Level 5 | 83% | — |
Knowledge GPT-5.4 mini leads
GPT-4.1: 37.1 (#160), GPT-5.4 mini: 51.5 (#67)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| GPQA Diamond | 66.9% | 86.9% |
| SimpleQA Verified | 31.1% | 29.4% |
| Vectara Hallucination Rate | 5.6% | 5.5% |
| LMArena Expert | 1364 | 1435 |
| Humanity's Last Exam | 5.4% | — |
| MMLU-Pro | 81.1% | — |
| GPQA (HELM) | 65.9% | — |
Multimodal GPT-5.4 mini leads
GPT-4.1: 38.2 (#67), GPT-5.4 mini: 39.7 (#56)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| LMArena Vision | 1211 | 1245 |
| GeoBench | 72% | — |
Multilingual GPT-5.4 mini leads
GPT-4.1: 49.4 (#133), GPT-5.4 mini: 51.9 (#96)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| LMArena Non-English | 1370 | 1405 |
| LMArena Chinese | 1382 | 1446 |
| LMArena French | 1382 | 1440 |
| LMArena German | 1381 | 1409 |
| LMArena Japanese | 1319 | 1374 |
| LMArena Korean | 1339 | 1368 |
| LMArena Russian | 1377 | 1417 |
| LMArena Spanish | 1376 | 1405 |
Instruction Following GPT-5.4 mini leads
GPT-4.1: 71.3 (#153), GPT-5.4 mini: 74.1 (#102)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| LMArena Instruction Following | 1367 | 1405 |
| IFEval | 83.8% | — |
Long Context GPT-5.4 mini leads
GPT-4.1: 40.0 (#163), GPT-5.4 mini: 43.0 (#112)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| LMArena Longer Query | 1385 | 1407 |
| Fiction.LiveBench | 63.9% | — |
Writing & Preference GPT-5.4 mini leads
GPT-4.1: 57.6 (#125), GPT-5.4 mini: 64.0 (#58)
| Benchmark | GPT-4.1 | GPT-5.4 mini |
|---|---|---|
| LMArena Text | 1383 | 1412 |
| LMArena Creative Writing | 1363 | 1370 |
| EQ-Bench Creative Writing | 1420 | 1665 |
| LMArena Multi-Turn | 1398 | 1429 |
| WildBench | 85.4% | — |
Frequently asked questions
Is GPT-4.1 better than GPT-5.4 mini?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 35.9 on the Noometry Index.
Which is cheaper, GPT-4.1 or GPT-5.4 mini?
GPT-5.4 mini is cheaper. It lists at $0.75 per million input tokens and $4.50 per million output tokens; GPT-4.1 lists at $2 and $8.
Is GPT-4.1 or GPT-5.4 mini better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
GPT-4.1 does, with 1.05M tokens against 400K.
How many benchmarks do GPT-4.1 and GPT-5.4 mini share?
36 benchmarks have published results for both models. GPT-4.1 has 52 scored results on Noometry and GPT-5.4 mini has 46.