Model comparison
gpt-oss-120b vs Phi 3 Mini 4k Instruct
gpt-oss-120b is the stronger model overall, scoring 36.0 to 27.4 on the Noometry Index.
Last verified . 18 shared benchmarks.
MicrosoftSummary
- They share 18 benchmarks with published results for both. gpt-oss-120b scores higher in 7 categories and Phi 3 Mini 4k Instruct in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 50.6 to 24.5.
- The biggest single-benchmark swing is Chess Puzzles: 20% for gpt-oss-120b and 0% for Phi 3 Mini 4k Instruct.
Side by side
| gpt-oss-120b | Phi 3 Mini 4k Instruct | |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 36.0 | 27.4 |
| Released | 2025-08-05 | 2024-04-23 |
| Weights | Open | Open |
| Context window | 131K | — |
| Max output | 41K | — |
| Input $ / M tokens | $0.037 | — |
| Output $ / M tokens | $0.17 | — |
| Results tracked | 48 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-120b leads
gpt-oss-120b: 33.2 (#251), Phi 3 Mini 4k Instruct: 26.1 (#319)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Coding | 1380 | 1093 |
| SWE-bench Verified (bash only) | 26% | — |
| Aider Polyglot | 41.8% | — |
| SciCode | 36% | — |
| WeirdML | 48.2% | — |
| LiveBench Coding | — | 15.5% |
| ALE-Bench | 575.62 | — |
| AlgoTune | 1.41 | — |
| HumanEval+ | — | 59.1% |
| MBPP+ | — | 54.2% |
Agentic & Tool Use Not comparable
gpt-oss-120b: 12.1 (#145), Phi 3 Mini 4k Instruct: —
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| Terminal-Bench | 18.7% | — |
| APEX-Agents | 4.4% | — |
| METR Time Horizons | 56.6% | — |
| Vending-Bench 2 | -21.53 | — |
Reasoning gpt-oss-120b leads
gpt-oss-120b: 19.5 (#236), Phi 3 Mini 4k Instruct: 13.6 (#326)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| Chess Puzzles | 20% | 0% |
| LMArena Hard Prompts | 1364 | 1072 |
| SimpleBench | 22.1% | — |
| Kagi LLM Benchmark | 58.6% | — |
| CritPt | 1.1% | — |
| LiveBench Reasoning | — | 26.8% |
| Mystery Game Puzzles | 2% | — |
| DTBench | 76.3% | — |
| LiveBench Data Analysis | — | 34.7% |
| LMCA | 22.1% | — |
| Surface Evolver Bench | 25% | — |
| Adversarial NLI | — | 52.8% |
| BIG-Bench Hard | — | 71.7% |
| Epoch Capabilities Index | 139.94 | — |
| HellaSwag | — | 76.7% |
| LiveBench | — | 22.4% |
| WinoGrande | — | 70.8% |
Math gpt-oss-120b leads
gpt-oss-120b: 50.6 (#43), Phi 3 Mini 4k Instruct: 24.5 (#258)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Math | 1388 | 1111 |
| OTIS Mock AIME 2024-2025 | 88.9% | — |
| Omni-MATH | 68.8% | — |
| LiveBench Math | — | 15.7% |
Knowledge gpt-oss-120b leads
gpt-oss-120b: 42.3 (#89), Phi 3 Mini 4k Instruct: 27.9 (#241)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Expert | 1356 | 1045 |
| GPQA Diamond | 75.8% | — |
| MMLU-Pro | 79.5% | — |
| Confabulations | 15.7% | — |
| Vectara Hallucination Rate | 14.2% | — |
| GPQA (HELM) | 68.4% | — |
| ARC (AI2) Challenge | — | 84.9% |
| MMLU | — | 68.8% |
| OpenBookQA | — | 88% |
| TriviaQA | — | 64% |
Multilingual gpt-oss-120b leads
gpt-oss-120b: 48.2 (#142), Phi 3 Mini 4k Instruct: 26.5 (#275)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Non-English | 1350 | 1021 |
| LMArena Chinese | 1384 | 1021 |
| LMArena French | 1370 | 1075 |
| LMArena German | 1355 | 1044 |
| LMArena Japanese | 1332 | 935 |
| LMArena Korean | 1283 | 905 |
| LMArena Russian | 1343 | 1022 |
| LMArena Spanish | 1388 | 1084 |
Instruction Following gpt-oss-120b leads
gpt-oss-120b: 69.5 (#168), Phi 3 Mini 4k Instruct: 48.0 (#298)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Instruction Following | 1318 | 1053 |
| LiveBench Instruction Following | — | 39.1% |
| IFEval | 83.6% | — |
Long Context Too close to call
gpt-oss-120b: 31.4 (#273), Phi 3 Mini 4k Instruct: 31.7 (#271)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Longer Query | 1319 | 1044 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference gpt-oss-120b leads
gpt-oss-120b: 46.7 (#213), Phi 3 Mini 4k Instruct: 27.7 (#295)
| Benchmark | gpt-oss-120b | Phi 3 Mini 4k Instruct |
|---|---|---|
| LMArena Text | 1366 | 1073 |
| LMArena Creative Writing | 1274 | 1037 |
| LMArena Multi-Turn | 1340 | 1018 |
| Short-Story Creative Writing | 77.1% | — |
| EQ-Bench Creative Writing | 961 | — |
| WildBench | 84.5% | — |
| LiveBench Language | — | 9.2% |
Frequently asked questions
Is gpt-oss-120b better than Phi 3 Mini 4k Instruct?
gpt-oss-120b is the stronger model overall, scoring 36.0 to 27.4 on the Noometry Index.
Is gpt-oss-120b or Phi 3 Mini 4k Instruct better for coding?
gpt-oss-120b scores higher on coding benchmarks: 33.2 versus 26.1 in the Noometry coding category.
How many benchmarks do gpt-oss-120b and Phi 3 Mini 4k Instruct share?
18 benchmarks have published results for both models. gpt-oss-120b has 48 scored results on Noometry and Phi 3 Mini 4k Instruct has 35.