Model comparison
Gemma 3 4B vs MiMo-V2.5
MiMo-V2.5 is the stronger model overall, scoring 43.0 to 27.9 on the Noometry Index. Gemma 3 4B costs 3.5× less per token, which makes it the better buy when MiMo-V2.5's lead doesn't matter for your workload.
Last verified . 12 shared benchmarks.
GoogleSummary
- They share 12 benchmarks with published results for both. Gemma 3 4B scores higher in 0 categories and MiMo-V2.5 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiMo-V2.5 leads 40.2 to 11.8.
- Gemma 3 4B is cheaper at $0.04 / $0.08 per million input/output tokens, against $0.14 / $0.28 for MiMo-V2.5.
- MiMo-V2.5 accepts more context: 1.05M tokens versus 131K.
Side by side
| Gemma 3 4B | MiMo-V2.5 | |
|---|---|---|
| Provider | Xiaomi | |
| Noometry Index | 27.9 | 43.0 |
| Released | 2025-03-12 | 2026-04-22 |
| Weights | Open | Open |
| Context window | 131K | 1.05M |
| Max output | 4K | 131K |
| Input $ / M tokens | $0.04 | $0.14 |
| Output $ / M tokens | $0.08 | $0.28 |
| Results tracked | 22 | 23 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiMo-V2.5 leads
Gemma 3 4B: 35.3 (#214), MiMo-V2.5: 43.6 (#75)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Coding | 1230 | 1469 |
| LMArena WebDev | — | 1438 |
| SciCode | — | 43.1% |
| ALE-Bench | — | 513.95 |
Agentic & Tool Use Not comparable
Gemma 3 4B: 20.7 (#134), MiMo-V2.5: —
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| Berkeley Function Calling Leaderboard | 19.6% | — |
Reasoning MiMo-V2.5 leads
Gemma 3 4B: 12.7 (#330), MiMo-V2.5: 27.8 (#98)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Hard Prompts | 1253 | 1450 |
| Kagi LLM Benchmark | 25.2% | — |
| CritPt | — | 3.7% |
| Chess Puzzles | 0% | — |
| DTBench | 50.9% | — |
| LMCA | 2.8% | — |
| Epoch Capabilities Index | 116.02 | — |
Math MiMo-V2.5 leads
Gemma 3 4B: 15.8 (#286), MiMo-V2.5: 35.4 (#145)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Math | 1239 | 1436 |
| OTIS Mock AIME 2024-2025 | 7.5% | — |
| ProofBench | — | 16% |
Knowledge MiMo-V2.5 leads
Gemma 3 4B: 11.8 (#294), MiMo-V2.5: 40.2 (#109)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Expert | 1223 | 1459 |
| GPQA Diamond | 23.2% | — |
| Vectara Hallucination Rate | 6.4% | — |
Multimodal Not comparable
Gemma 3 4B: —, MiMo-V2.5: 39.0 (#53)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Vision | — | 1245 |
Multilingual MiMo-V2.5 leads
Gemma 3 4B: 42.7 (#189), MiMo-V2.5: 52.0 (#96)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Non-English | 1273 | 1404 |
| LMArena German | 1281 | 1421 |
| LMArena Russian | 1294 | 1395 |
| LMArena Chinese | — | 1467 |
| LMArena French | — | 1447 |
| LMArena Japanese | — | 1304 |
| LMArena Korean | — | 1363 |
| LMArena Spanish | — | 1412 |
Instruction Following MiMo-V2.5 leads
Gemma 3 4B: 65.4 (#219), MiMo-V2.5: 75.7 (#55)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Instruction Following | 1239 | 1433 |
Long Context MiMo-V2.5 leads
Gemma 3 4B: 38.6 (#189), MiMo-V2.5: 44.1 (#69)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Longer Query | 1274 | 1444 |
Writing & Preference MiMo-V2.5 leads
Gemma 3 4B: 42.2 (#234), MiMo-V2.5: 61.8 (#82)
| Benchmark | Gemma 3 4B | MiMo-V2.5 |
|---|---|---|
| LMArena Text | 1291 | 1427 |
| LMArena Creative Writing | 1270 | 1392 |
| LMArena Multi-Turn | 1255 | 1445 |
| EQ-Bench Creative Writing | 1068 | — |
Frequently asked questions
Is Gemma 3 4B better than MiMo-V2.5?
MiMo-V2.5 is the stronger model overall, scoring 43.0 to 27.9 on the Noometry Index. Gemma 3 4B costs 3.5× less per token, which makes it the better buy when MiMo-V2.5's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 4B or MiMo-V2.5?
Gemma 3 4B is cheaper. It lists at $0.04 per million input tokens and $0.08 per million output tokens; MiMo-V2.5 lists at $0.14 and $0.28.
Is Gemma 3 4B or MiMo-V2.5 better for coding?
MiMo-V2.5 scores higher on coding benchmarks: 43.6 versus 35.3 in the Noometry coding category.
Which has the bigger context window?
MiMo-V2.5 does, with 1.05M tokens against 131K.
How many benchmarks do Gemma 3 4B and MiMo-V2.5 share?
12 benchmarks have published results for both models. Gemma 3 4B has 22 scored results on Noometry and MiMo-V2.5 has 23.