Model comparison
GPT-5.4 mini vs Phi 3 Mini 4k Instruct
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 27.9 on the Noometry Index.
Last verified . 18 shared benchmarks.
MicrosoftSummary
- They share 18 benchmarks with published results for both. GPT-5.4 mini scores higher in 8 categories and Phi 3 Mini 4k Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-5.4 mini leads 64.0 to 27.6.
- The biggest single-benchmark swing is Chess Puzzles: 24% for GPT-5.4 mini and 0% for Phi 3 Mini 4k Instruct.
- Phi 3 Mini 4k Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 mini | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 45.0 | 27.9 |
| Released | 2026-03-17 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | 400K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.75 | — |
| Output $ / M tokens | $4.50 | — |
| Results tracked | 46 | 35 |
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), Phi 3 Mini 4k Instruct: 26.6 (#323)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1438 | 1093 |
| FrontierCode | 27% | — |
| LMArena WebDev | 1397 | — |
| SciCode | 49.9% | — |
| WeirdML | 60.3% | — |
| LiveBench Coding | — | 15.5% |
| ALE-Bench | 1,189 | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
GPT-5.4 mini: 29.9 (#81), Phi 3 Mini 4k Instruct: —
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| DeepResearch Bench | 36.3% | — |
Reasoning GPT-5.4 mini leads
GPT-5.4 mini: 30.4 (#85), Phi 3 Mini 4k Instruct: 14.1 (#328)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 24% | 0% |
| LMArena Hard Prompts | 1424 | 1072 |
| ARC-AGI-2 | 18.9% | — |
| Kagi LLM Benchmark | 37.9% | — |
| NYT Connections (extended) | 61.8% | — |
| ARC-AGI-1 | 63.7% | — |
| CritPt | 10% | — |
| Thematic Generalization | 61.7% | — |
| LiveBench Reasoning | — | 26.8% |
| Mystery Game Puzzles | 11% | — |
| DTBench | 80% | — |
| LiveBench Data Analysis | — | 34.7% |
| LMCA | 40.8% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 148.84 | — |
| ForecastBench | 57 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
| WinoGrande | — | 70.8% |
Math GPT-5.4 mini leads
GPT-5.4 mini: 45.5 (#75), Phi 3 Mini 4k Instruct: 26.6 (#257)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1419 | 1111 |
| FrontierMath (Tiers 1-3) | 51.2% | — |
| FrontierMath Tier 4 | 9.8% | — |
| OTIS Mock AIME 2024-2025 | 88.9% | — |
| ProofBench | 21% | — |
| LiveBench Math | — | 15.7% |
| FrontierMath (Feb 2025 set) | 28.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5.4 mini leads
GPT-5.4 mini: 51.5 (#67), Phi 3 Mini 4k Instruct: 28.5 (#246)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1435 | 1045 |
| GPQA Diamond | 86.9% | — |
| SimpleQA Verified | 29.4% | — |
| Vectara Hallucination Rate | 5.5% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multimodal Not comparable
GPT-5.4 mini: 39.7 (#56), Phi 3 Mini 4k Instruct: —
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Vision | 1245 | — |
Multilingual GPT-5.4 mini leads
GPT-5.4 mini: 51.9 (#96), Phi 3 Mini 4k Instruct: 26.3 (#280)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1405 | 1021 |
| LMArena Chinese | 1446 | 1021 |
| LMArena French | 1440 | 1076 |
| LMArena German | 1409 | 1044 |
| LMArena Japanese | 1374 | 935 |
| LMArena Korean | 1368 | 905 |
| LMArena Russian | 1417 | 1022 |
| LMArena Spanish | 1405 | 1085 |
Instruction Following GPT-5.4 mini leads
GPT-5.4 mini: 74.1 (#102), Phi 3 Mini 4k Instruct: 47.7 (#303)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1405 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
Long Context GPT-5.4 mini leads
GPT-5.4 mini: 43.0 (#112), Phi 3 Mini 4k Instruct: 31.7 (#276)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1407 | 1044 |
Writing & Preference GPT-5.4 mini leads
GPT-5.4 mini: 64.0 (#58), Phi 3 Mini 4k Instruct: 27.6 (#300)
| Benchmark | GPT-5.4 mini | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1412 | 1073 |
| LMArena Creative Writing | 1370 | 1037 |
| LMArena Multi-Turn | 1429 | 1018 |
| EQ-Bench Creative Writing | 1665 | — |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is GPT-5.4 mini better than Phi 3 Mini 4k Instruct?
GPT-5.4 mini is the stronger model overall, scoring 45.0 to 27.9 on the Noometry Index.
Is GPT-5.4 mini or Phi 3 Mini 4k Instruct better for coding?
GPT-5.4 mini scores higher on coding benchmarks: 45.2 versus 26.6 in the Noometry coding category.
How many benchmarks do GPT-5.4 mini and Phi 3 Mini 4k Instruct share?
18 benchmarks have published results for both models. GPT-5.4 mini has 46 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.