Model comparison
MiniMax-M2 vs Phi-4
MiniMax-M2 is the stronger model overall, scoring 37.4 to 31.2 on the Noometry Index. Phi-4 costs 6.0× less per token, which makes it the better buy when MiniMax-M2's lead doesn't matter for your workload.
Last verified . 15 shared benchmarks.
MiniMax
MicrosoftSummary
- They share 15 benchmarks with published results for both. MiniMax-M2 scores higher in 9 categories and Phi-4 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where MiniMax-M2 leads 37.3 to 20.8.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.
- MiniMax-M2 accepts more context: 205K tokens versus 128K.
Side by side
| MiniMax-M2 | Phi-4 | |
|---|---|---|
| Provider | MiniMax | Microsoft |
| Noometry Index | 37.4 | 31.2 |
| Released | 2025-10-27 | 2024-12-11 |
| Weights | Open | Open |
| Context window | 205K | 128K |
| Max output | 131K | 4K |
| Input $ / M tokens | $0.30 | $0.07 |
| Output $ / M tokens | $1.20 | $0.14 |
| Results tracked | 21 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2 leads
MiniMax-M2: 39.3 (#159), Phi-4: 34.4 (#239)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Coding | 1370 | 1231 |
| SWE-bench Verified (bash only) | 61% | — |
| LMArena WebDev | 1297 | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
Agentic & Tool Use MiniMax-M2 leads
MiniMax-M2: 25.1 (#109), Phi-4: 22.8 (#128)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| Terminal-Bench | 30% | — |
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |
| Vending-Bench 2 | 160.6 | — |
Reasoning MiniMax-M2 leads
MiniMax-M2: 19.4 (#258), Phi-4: 17.7 (#291)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1357 | 1220 |
| Kagi LLM Benchmark | 57.8% | — |
| NYT Connections (extended) | 14.8% | — |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 47.8% |
| LiveBench Data Analysis | — | 45.2% |
| Epoch Capabilities Index | — | 130.42 |
| LiveBench | — | 41.6% |
Math MiniMax-M2 leads
MiniMax-M2: 37.3 (#160), Phi-4: 20.8 (#285)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Math | 1352 | 1246 |
| OTIS Mock AIME 2024-2025 | — | 13.8% |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge MiniMax-M2 leads
MiniMax-M2: 37.0 (#163), Phi-4: 32.6 (#209)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Expert | 1337 | 1203 |
| GPQA Diamond | — | 56.1% |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |
Multilingual MiniMax-M2 leads
MiniMax-M2: 45.3 (#171), Phi-4: 37.2 (#237)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Non-English | 1313 | 1197 |
| LMArena Chinese | 1366 | 1212 |
| LMArena French | 1335 | 1224 |
| LMArena German | 1355 | 1222 |
| LMArena Russian | 1331 | 1209 |
| LMArena Spanish | 1326 | 1234 |
| LMArena Japanese | — | 1158 |
| LMArena Korean | — | 1151 |
Instruction Following MiniMax-M2 leads
MiniMax-M2: 70.2 (#166), Phi-4: 60.4 (#251)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1328 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context MiniMax-M2 leads
MiniMax-M2: 40.5 (#153), Phi-4: 36.9 (#226)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1331 | 1217 |
Writing & Preference MiniMax-M2 leads
MiniMax-M2: 53.0 (#162), Phi-4: 40.5 (#244)
| Benchmark | MiniMax-M2 | Phi-4 |
|---|---|---|
| LMArena Text | 1340 | 1217 |
| LMArena Creative Writing | 1286 | 1182 |
| LMArena Multi-Turn | 1361 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is MiniMax-M2 better than Phi-4?
MiniMax-M2 is the stronger model overall, scoring 37.4 to 31.2 on the Noometry Index. Phi-4 costs 6.0× less per token, which makes it the better buy when MiniMax-M2's lead doesn't matter for your workload.
Which is cheaper, MiniMax-M2 or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; MiniMax-M2 lists at $0.30 and $1.20.
Is MiniMax-M2 or Phi-4 better for coding?
MiniMax-M2 scores higher on coding benchmarks: 39.3 versus 34.4 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2 does, with 205K tokens against 128K.
How many benchmarks do MiniMax-M2 and Phi-4 share?
15 benchmarks have published results for both models. MiniMax-M2 has 21 scored results on Noometry and Phi-4 has 37.