Model comparison
GPT-5.4 nano vs Phi 3 Mini 128k Instruct
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 29.7 on the Noometry Index.
Last verified . 17 shared benchmarks.
MicrosoftSummary
- They share 17 benchmarks with published results for both. GPT-5.4 nano scores higher in 8 categories and Phi 3 Mini 128k Instruct in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where GPT-5.4 nano leads 55.7 to 27.1.
- Phi 3 Mini 128k Instruct has downloadable open weights; the other is API-only.
Side by side
| GPT-5.4 nano | Phi 3 Mini 128k Instruct | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 41.9 | 29.7 |
| Released | 2026-03-17 | 2024-04-23 |
| Weights | Proprietary | Open |
| Context window | 400K | — |
| Max output | 128K | — |
| Input $ / M tokens | $0.20 | — |
| Output $ / M tokens | $1.25 | — |
| Results tracked | 40 | 19 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 nano leads
GPT-5.4 nano: 43.6 (#84), Phi 3 Mini 128k Instruct: 28.8 (#312)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Coding | 1405 | 1039 |
| SciCode | 46.9% | — |
| WeirdML | 49.2% | — |
| BigCodeBench Instruct | — | 29.6% |
| BigCodeBench Complete | — | 40.6% |
| ALE-Bench | 1,005 | — |
Reasoning GPT-5.4 nano leads
GPT-5.4 nano: 23.7 (#173), Phi 3 Mini 128k Instruct: 19.6 (#256)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Hard Prompts | 1381 | 1028 |
| ARC-AGI-2 | 5.7% | — |
| Kagi LLM Benchmark | 39.7% | — |
| ARC-AGI-1 | 51.5% | — |
| CritPt | 9.3% | — |
| Chess Puzzles | 30% | — |
| Mystery Game Puzzles | 9% | — |
| DTBench | 80.3% | — |
| LMCA | 36.9% | — |
| Epoch Capabilities Index | 145.81 | — |
| ForecastBench | 57.3 | — |
Math GPT-5.4 nano leads
GPT-5.4 nano: 40.9 (#88), Phi 3 Mini 128k Instruct: 31.6 (#222)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Math | 1406 | 1089 |
| FrontierMath (Tiers 1-3) | 44.9% | — |
| FrontierMath Tier 4 | 12.2% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| ProofBench | 5% | — |
| FrontierMath (Feb 2025 set) | 25.9% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge GPT-5.4 nano leads
GPT-5.4 nano: 41.9 (#103), Phi 3 Mini 128k Instruct: 26.8 (#254)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Expert | 1396 | 984 |
| GPQA Diamond | 78.5% | — |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 3.1% | — |
Multimodal Not comparable
GPT-5.4 nano: 36.7 (#78), Phi 3 Mini 128k Instruct: —
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Vision | 1196 | — |
Multilingual GPT-5.4 nano leads
GPT-5.4 nano: 48.6 (#140), Phi 3 Mini 128k Instruct: 25.2 (#285)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Non-English | 1359 | 1000 |
| LMArena Chinese | 1392 | 1016 |
| LMArena French | 1396 | 1039 |
| LMArena German | 1367 | 1006 |
| LMArena Japanese | 1343 | 899 |
| LMArena Korean | 1320 | 856 |
| LMArena Russian | 1363 | 1004 |
| LMArena Spanish | 1371 | 1059 |
Instruction Following GPT-5.4 nano leads
GPT-5.4 nano: 71.9 (#144), Phi 3 Mini 128k Instruct: 51.8 (#294)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Instruction Following | 1362 | 1023 |
Long Context GPT-5.4 nano leads
GPT-5.4 nano: 41.6 (#137), Phi 3 Mini 128k Instruct: 30.4 (#289)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Longer Query | 1366 | 996 |
Writing & Preference GPT-5.4 nano leads
GPT-5.4 nano: 55.7 (#142), Phi 3 Mini 128k Instruct: 27.1 (#301)
| Benchmark | GPT-5.4 nano | Phi 3 Mini 128k Instruct |
|---|---|---|
| LMArena Text | 1372 | 1050 |
| LMArena Creative Writing | 1314 | 1024 |
| LMArena Multi-Turn | 1382 | 989 |
Frequently asked questions
Is GPT-5.4 nano better than Phi 3 Mini 128k Instruct?
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 29.7 on the Noometry Index.
Is GPT-5.4 nano or Phi 3 Mini 128k Instruct better for coding?
GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 28.8 in the Noometry coding category.
How many benchmarks do GPT-5.4 nano and Phi 3 Mini 128k Instruct share?
17 benchmarks have published results for both models. GPT-5.4 nano has 40 scored results on Noometry and Phi 3 Mini 128k Instruct has 19.