Model comparison
Gemma 3 27B vs MiniMax-M2.5
MiniMax-M2.5 is the stronger model overall, scoring 38.3 to 30.8 on the Noometry Index. Gemma 3 27B costs 5.2× less per token, which makes it the better buy when MiniMax-M2.5's lead doesn't matter for your workload.
Last verified . 21 shared benchmarks.
Google
MiniMaxSummary
- They share 21 benchmarks with published results for both. Gemma 3 27B scores higher in 0 categories and MiniMax-M2.5 in 9 categories; 6 gaps are clear of the uncertainty.
- The widest gap is in coding, where MiniMax-M2.5 leads 48.1 to 22.5.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 40.4% for Gemma 3 27B and 55.2% for MiniMax-M2.5.
- Gemma 3 27B is cheaper at $0.08 / $0.16 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.5.
- MiniMax-M2.5 accepts more context: 205K tokens versus 131K.
Side by side
| Gemma 3 27B | MiniMax-M2.5 | |
|---|---|---|
| Provider | MiniMax | |
| Noometry Index | 30.8 | 38.3 |
| Released | 2025-03-11 | 2026-02-12 |
| Weights | Open | Open |
| Context window | 131K | 205K |
| Max output | 8K | 131K |
| Input $ / M tokens | $0.08 | $0.30 |
| Output $ / M tokens | $0.16 | $1.20 |
| Results tracked | 43 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
Gemma 3 27B: 22.5 (#334), MiniMax-M2.5: 48.1 (#58)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Coding | 1322 | 1381 |
| SWE-bench Verified (bash only) | — | 75.8% |
| Aider Polyglot | 4.9% | — |
| LMArena WebDev | — | 1387 |
| SWE-bench Multilingual | — | 68.3% |
| SciCode | 21.2% | — |
| LiveBench Coding | 39.9% | — |
| ALE-Bench | — | 618.17 |
Agentic & Tool Use MiniMax-M2.5 leads
Gemma 3 27B: 25.1 (#110), MiniMax-M2.5: 30.4 (#77)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | — | 42.7% |
| Berkeley Function Calling Leaderboard | 29.5% | — |
| Vending-Bench 2 | — | -23.16 |
Reasoning Too close to call
Gemma 3 27B: 16.7 (#301), MiniMax-M2.5: 17.5 (#292)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| Kagi LLM Benchmark | 40.4% | 55.2% |
| LMArena Hard Prompts | 1340 | 1372 |
| Epoch Capabilities Index | 130.04 | 146.68 |
| ARC-AGI-2 | — | 4.9% |
| NYT Connections (extended) | — | 16.8% |
| ARC-AGI-1 | — | 63.7% |
| CritPt | 0% | — |
| Chess Puzzles | 0% | — |
| LiveBench Reasoning | 43.8% | — |
| DTBench | 52.5% | — |
| LiveBench Data Analysis | 51.5% | — |
| LMCA | 12.3% | — |
| LiveBench | 50% | — |
Math MiniMax-M2.5 leads
Gemma 3 27B: 25.9 (#265), MiniMax-M2.5: 26.9 (#253)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Math | 1312 | 1378 |
| OTIS Mock AIME 2024-2025 | 22.5% | — |
| ProofBench | — | 4% |
| LiveBench Math | 55.4% | — |
| MATH Level 5 | 74% | — |
Knowledge MiniMax-M2.5 leads
Gemma 3 27B: 25.5 (#261), MiniMax-M2.5: 39.2 (#135)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| Vectara Hallucination Rate | 7.4% | 9.1% |
| LMArena Expert | 1304 | 1379 |
| GPQA Diamond | 47.7% | — |
| Confabulations | 40.3% | — |
Multimodal Not comparable
Gemma 3 27B: 32.6 (#100), MiniMax-M2.5: —
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Vision | 1164 | — |
| GeoBench | 52% | — |
Multilingual Too close to call
Gemma 3 27B: 46.9 (#155), MiniMax-M2.5: 47.1 (#152)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1334 | 1338 |
| LMArena Chinese | 1346 | 1393 |
| LMArena French | 1368 | 1362 |
| LMArena German | 1362 | 1362 |
| LMArena Japanese | 1287 | 1171 |
| LMArena Korean | 1308 | 1232 |
| LMArena Russian | 1349 | 1358 |
| LMArena Spanish | 1349 | 1354 |
Instruction Following Too close to call
Gemma 3 27B: 70.6 (#160), MiniMax-M2.5: 71.5 (#148)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1321 | 1353 |
| LiveBench Instruction Following | 74.9% | — |
Long Context MiniMax-M2.5 leads
Gemma 3 27B: 27.6 (#293), MiniMax-M2.5: 37.5 (#216)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1333 | 1366 |
| Fiction.LiveBench | 33.3% | — |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
Writing & Preference MiniMax-M2.5 leads
Gemma 3 27B: 52.5 (#168), MiniMax-M2.5: 53.9 (#153)
| Benchmark | Gemma 3 27B | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1358 | 1359 |
| LMArena Creative Writing | 1346 | 1331 |
| EQ-Bench Creative Writing | 1266 | 1361 |
| LMArena Multi-Turn | 1345 | 1364 |
| Short-Story Creative Writing | 79.9% | — |
| LiveBench Language | 34.6% | — |
Frequently asked questions
Is Gemma 3 27B better than MiniMax-M2.5?
MiniMax-M2.5 is the stronger model overall, scoring 38.3 to 30.8 on the Noometry Index. Gemma 3 27B costs 5.2× less per token, which makes it the better buy when MiniMax-M2.5's lead doesn't matter for your workload.
Which is cheaper, Gemma 3 27B or MiniMax-M2.5?
Gemma 3 27B is cheaper. It lists at $0.08 per million input tokens and $0.16 per million output tokens; MiniMax-M2.5 lists at $0.30 and $1.20.
Is Gemma 3 27B or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 48.1 versus 22.5 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.5 does, with 205K tokens against 131K.
How many benchmarks do Gemma 3 27B and MiniMax-M2.5 share?
21 benchmarks have published results for both models. Gemma 3 27B has 43 scored results on Noometry and MiniMax-M2.5 has 33.