Model comparison
Deepseek Coder v2 vs MiniMax-M2
MiniMax-M2 is the stronger model overall, scoring 37.4 to 35.9 on the Noometry Index.
Last verified . 15 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 15 benchmarks with published results for both. Deepseek Coder v2 scores higher in 1 category and MiniMax-M2 in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where MiniMax-M2 leads 53.0 to 38.2.
Side by side
| Deepseek Coder v2 | MiniMax-M2 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 35.9 | 37.4 |
| Released | 2024-06-17 | 2025-10-27 |
| Weights | Open | Open |
| Context window | — | 205K |
| Max output | — | 131K |
| Input $ / M tokens | — | $0.30 |
| Output $ / M tokens | — | $1.20 |
| Results tracked | 24 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2 leads
Deepseek Coder v2: 38.1 (#183), MiniMax-M2: 39.3 (#159)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Coding | 1251 | 1370 |
| SWE-bench Verified (bash only) | — | 61% |
| LMArena WebDev | — | 1297 |
| BigCodeBench Instruct | 48.2% | — |
| BigCodeBench Complete | 59.7% | — |
| HumanEval+ | 82.3% | — |
| MBPP+ | 75.1% | — |
Agentic & Tool Use Not comparable
Deepseek Coder v2: —, MiniMax-M2: 25.1 (#109)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| Terminal-Bench | — | 30% |
| Vending-Bench 2 | — | 160.6 |
Reasoning Deepseek Coder v2 leads
Deepseek Coder v2: 23.6 (#176), MiniMax-M2: 19.4 (#258)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Hard Prompts | 1207 | 1357 |
| Kagi LLM Benchmark | — | 57.8% |
| NYT Connections (extended) | — | 14.8% |
| WinoGrande | 83.7% | — |
Math MiniMax-M2 leads
Deepseek Coder v2: 34.9 (#190), MiniMax-M2: 37.3 (#160)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Math | 1241 | 1352 |
| GSM8K | 94.5% | — |
Knowledge MiniMax-M2 leads
Deepseek Coder v2: 32.3 (#212), MiniMax-M2: 37.0 (#163)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Expert | 1181 | 1337 |
| ARC (AI2) Challenge | 64.3% | — |
Multilingual MiniMax-M2 leads
Deepseek Coder v2: 36.3 (#240), MiniMax-M2: 45.3 (#171)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Non-English | 1182 | 1313 |
| LMArena Chinese | 1201 | 1366 |
| LMArena French | 1185 | 1335 |
| LMArena German | 1164 | 1355 |
| LMArena Russian | 1188 | 1331 |
| LMArena Spanish | 1153 | 1326 |
| LMArena Japanese | 1126 | — |
| LMArena Korean | 1104 | — |
Instruction Following MiniMax-M2 leads
Deepseek Coder v2: 61.7 (#242), MiniMax-M2: 70.2 (#166)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Instruction Following | 1180 | 1328 |
Long Context MiniMax-M2 leads
Deepseek Coder v2: 37.0 (#224), MiniMax-M2: 40.5 (#153)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Longer Query | 1219 | 1331 |
Writing & Preference MiniMax-M2 leads
Deepseek Coder v2: 38.2 (#253), MiniMax-M2: 53.0 (#162)
| Benchmark | Deepseek Coder v2 | MiniMax-M2 |
|---|---|---|
| LMArena Text | 1191 | 1340 |
| LMArena Creative Writing | 1120 | 1286 |
| LMArena Multi-Turn | 1177 | 1361 |
Frequently asked questions
Is Deepseek Coder v2 better than MiniMax-M2?
MiniMax-M2 is the stronger model overall, scoring 37.4 to 35.9 on the Noometry Index.
Is Deepseek Coder v2 or MiniMax-M2 better for coding?
MiniMax-M2 scores higher on coding benchmarks: 39.3 versus 38.1 in the Noometry coding category.
How many benchmarks do Deepseek Coder v2 and MiniMax-M2 share?
15 benchmarks have published results for both models. Deepseek Coder v2 has 24 scored results on Noometry and MiniMax-M2 has 21.