Model comparison
MiMo-V2.5-Pro vs Phi-4
MiMo-V2.5-Pro is the stronger model overall, scoring 44.8 to 30.9 on the Noometry Index. Phi-4 costs 6.2× less per token, which makes it the better buy when MiMo-V2.5-Pro's lead doesn't matter for your workload.
Last verified . 17 shared benchmarks.
MicrosoftSummary
- They share 17 benchmarks with published results for both. MiMo-V2.5-Pro scores higher in 8 categories and Phi-4 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where MiMo-V2.5-Pro leads 65.5 to 40.7.
- Phi-4 is cheaper at $0.07 / $0.14 per million input/output tokens, against $0.43 / $0.87 for MiMo-V2.5-Pro.
- MiMo-V2.5-Pro accepts more context: 1.05M tokens versus 128K.
Side by side
| MiMo-V2.5-Pro | Phi-4 | |
|---|---|---|
| Provider | Xiaomi | Microsoft |
| Noometry Index | 44.8 | 30.9 |
| Released | 2026-04-22 | 2024-12-11 |
| Weights | Open | Open |
| Context window | 1.05M | 128K |
| Max output | 131K | 4K |
| Input $ / M tokens | $0.43 | $0.07 |
| Output $ / M tokens | $0.87 | $0.14 |
| Results tracked | 27 | 37 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 47.0 (#58), Phi-4: 33.8 (#237)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Coding | 1502 | 1231 |
| LMArena WebDev | 1477 | — |
| SciCode | 50.2% | — |
| BigCodeBench Instruct | — | 45.5% |
| LiveBench Coding | — | 30.7% |
| BigCodeBench Complete | — | 55.4% |
| ALE-Bench | 899.8 | — |
Agentic & Tool Use Not comparable
MiMo-V2.5-Pro: —, Phi-4: 22.6 (#120)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| Berkeley Function Calling Leaderboard | — | 28.8% |
| BALROG | — | 11.6% |
Reasoning MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 25.9 (#124), Phi-4: 17.0 (#287)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Hard Prompts | 1487 | 1220 |
| NYT Connections (extended) | 34.4% | — |
| CritPt | 4% | — |
| Chess Puzzles | — | 1% |
| LiveBench Reasoning | — | 47.8% |
| DTBench | 84.5% | — |
| LiveBench Data Analysis | — | 45.2% |
| LMCA | 29.5% | — |
| Epoch Capabilities Index | — | 130.43 |
| LiveBench | — | 41.6% |
Math MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 38.6 (#84), Phi-4: 19.6 (#279)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Math | 1480 | 1246 |
| OTIS Mock AIME 2024-2025 | — | 13.8% |
| ProofBench | 22% | — |
| LiveBench Math | — | 42% |
| MATH Level 5 | — | 64.9% |
Knowledge MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 41.5 (#100), Phi-4: 32.5 (#204)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Expert | 1501 | 1203 |
| GPQA Diamond | — | 56.1% |
| Confabulations | — | 29.4% |
| Vectara Hallucination Rate | — | 3.7% |
| MMLU | — | 84.8% |
Multilingual MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 55.3 (#33), Phi-4: 37.4 (#232)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Non-English | 1449 | 1197 |
| LMArena Chinese | 1507 | 1211 |
| LMArena French | 1489 | 1223 |
| LMArena German | 1455 | 1222 |
| LMArena Japanese | 1418 | 1158 |
| LMArena Korean | 1436 | 1151 |
| LMArena Russian | 1450 | 1209 |
| LMArena Spanish | 1470 | 1234 |
Instruction Following MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 77.6 (#21), Phi-4: 60.6 (#246)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Instruction Following | 1477 | 1201 |
| LiveBench Instruction Following | — | 58.4% |
Long Context MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 45.4 (#38), Phi-4: 36.9 (#221)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Longer Query | 1482 | 1217 |
Writing & Preference MiMo-V2.5-Pro leads
MiMo-V2.5-Pro: 65.5 (#47), Phi-4: 40.7 (#239)
| Benchmark | MiMo-V2.5-Pro | Phi-4 |
|---|---|---|
| LMArena Text | 1465 | 1217 |
| LMArena Creative Writing | 1440 | 1182 |
| LMArena Multi-Turn | 1477 | 1206 |
| Short-Story Creative Writing | — | 62.6% |
| EQ-Bench Creative Writing | 1493 | — |
| EQ-Bench 4 | 1208 | — |
| LiveBench Language | — | 25.6% |
Frequently asked questions
Is MiMo-V2.5-Pro better than Phi-4?
MiMo-V2.5-Pro is the stronger model overall, scoring 44.8 to 30.9 on the Noometry Index. Phi-4 costs 6.2× less per token, which makes it the better buy when MiMo-V2.5-Pro's lead doesn't matter for your workload.
Which is cheaper, MiMo-V2.5-Pro or Phi-4?
Phi-4 is cheaper. It lists at $0.07 per million input tokens and $0.14 per million output tokens; MiMo-V2.5-Pro lists at $0.43 and $0.87.
Is MiMo-V2.5-Pro or Phi-4 better for coding?
MiMo-V2.5-Pro scores higher on coding benchmarks: 47.0 versus 33.8 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.5-Pro does, with 1.05M tokens against 128K.
How many benchmarks do MiMo-V2.5-Pro and Phi-4 share?
17 benchmarks have published results for both models. MiMo-V2.5-Pro has 27 scored results on Noometry and Phi-4 has 37.