Model comparison
Mistral Large vs Phi-4
Mistral Large and Phi-4 score almost the same on the Noometry Index (31.9 vs 31.2), so choose on price, context window or the category you care about most.
Last verified . 35 shared benchmarks.
Mistral AI
MicrosoftSummary
- They share 35 benchmarks with published results for both. Mistral Large scores higher in 5 categories and Phi-4 in 4 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in instruction following, where Mistral Large leads 67.9 to 60.4.
- The biggest single-benchmark swing is BigCodeBench Complete: 38.3% for Mistral Large and 55.4% for Phi-4.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $2 / $6 for Mistral Large.
- Mistral Large accepts more context: 131K tokens versus 128K.
Side by side
| Mistral Large | Phi-4 | |
|---|---|---|
| Provider | Mistral AI | Microsoft |
| Noometry Index | 31.9 | 31.2 |
| Released | 2024-02-26 | 2024-12-11 |
| Weights | Open | Open |
| Context window | 131K | 128K |
| Max output | 16K | 4K |
| Input $ / M tokens | $2 | $0.07 |
| Output $ / M tokens | $6 | $0.14 |
| Results tracked | 51 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Mistral Large: 34.3 (#240), Phi-4: 34.4 (#239)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| BigCodeBench Instruct | 30% | 45.5% |
| LiveBench Coding | 47.1% | 30.7% |
| LMArena Coding | 1277 | 1231 |
| BigCodeBench Complete | 38.3% | 55.4% |
| SciCode | 36.2% | — |
| ALE-Bench | 264.7 | — |
| HumanEval+ | 62.2% | — |
| MBPP+ | 59.5% | — |
Agentic & Tool Use Mistral Large leads
Mistral Large: 28.6 (#89), Phi-4: 22.8 (#128)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 38.4% | 28.8% |
| BALROG | — | 11.6% |
Reasoning Phi-4 leads
Mistral Large: 15.8 (#310), Phi-4: 17.7 (#291)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| LiveBench Reasoning | 43.5% | 47.8% |
| LMArena Hard Prompts | 1257 | 1220 |
| LiveBench Data Analysis | 50.1% | 45.2% |
| Epoch Capabilities Index | 128.52 | 130.42 |
| LiveBench | 48.4% | 41.6% |
| SimpleBench | 22.5% | — |
| CritPt | 0% | — |
| Chess Puzzles | — | 1% |
| DTBench | 65.1% | — |
| LMCA | 16.7% | — |
| ForecastBench | 57.1 | — |
Math Phi-4 leads
Mistral Large: 18.2 (#291), Phi-4: 20.8 (#285)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 8.5% | 13.8% |
| LiveBench Math | 42.5% | 42% |
| LMArena Math | 1262 | 1246 |
| MATH Level 5 | 50.3% | 64.9% |
| Omni-MATH | 28.1% | — |
| FrontierMath (Feb 2025 set) | 0.3% | — |
Knowledge Phi-4 leads
Mistral Large: 30.1 (#230), Phi-4: 32.6 (#209)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| GPQA Diamond | 51.3% | 56.1% |
| Confabulations | 21.4% | 29.4% |
| Vectara Hallucination Rate | 4.5% | 3.7% |
| LMArena Expert | 1232 | 1203 |
| MMLU | 80% | 84.8% |
| MMLU-Pro | 59.9% | — |
| GPQA (HELM) | 43.5% | — |
Multilingual Mistral Large leads
Mistral Large: 40.0 (#219), Phi-4: 37.2 (#237)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| LMArena Non-English | 1237 | 1197 |
| LMArena Chinese | 1240 | 1212 |
| LMArena French | 1325 | 1224 |
| LMArena German | 1254 | 1222 |
| LMArena Japanese | 1188 | 1158 |
| LMArena Korean | 1202 | 1151 |
| LMArena Russian | 1257 | 1209 |
| LMArena Spanish | 1268 | 1234 |
Instruction Following Mistral Large leads
Mistral Large: 67.9 (#191), Phi-4: 60.4 (#251)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| LiveBench Instruction Following | 67.9% | 58.4% |
| LMArena Instruction Following | 1249 | 1201 |
| IFEval | 87.7% | — |
Long Context Mistral Large leads
Mistral Large: 38.3 (#199), Phi-4: 36.9 (#226)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1261 | 1217 |
Writing & Preference Too close to call
Mistral Large: 40.7 (#242), Phi-4: 40.5 (#244)
| Benchmark | Mistral Large | Phi-4 |
|---|---|---|
| LMArena Text | 1266 | 1217 |
| LMArena Creative Writing | 1243 | 1182 |
| Short-Story Creative Writing | 69% | 62.6% |
| LMArena Multi-Turn | 1260 | 1206 |
| LiveBench Language | 39.4% | 25.6% |
| EQ-Bench Creative Writing | 985 | — |
| WildBench | 80.1% | — |
Frequently asked questions
Is Mistral Large better than Phi-4?
Mistral Large and Phi-4 score almost the same on the Noometry Index (31.9 vs 31.2), so choose on price, context window or the category you care about most.
Which is cheaper, Mistral Large or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; Mistral Large lists at $2 and $6.
Is Mistral Large or Phi-4 better for coding?
They score almost the same on coding (34.3 vs 34.4); test both on your own repository before choosing.
Which has the bigger context window?
Mistral Large does, with 131K tokens against 128K.
How many benchmarks do Mistral Large and Phi-4 share?
35 benchmarks have published results for both models. Mistral Large has 51 scored results on Noometry and Phi-4 has 37.