Model comparison
DeepSeek-V2.5 (Sep 2024) vs MiniMax-M3
MiniMax-M3 is the stronger model overall, scoring 43.8 to 37.6 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 17 benchmarks with published results for both. DeepSeek-V2.5 (Sep 2024) scores higher in 0 categories and MiniMax-M3 in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 34.8.
Side by side
| DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 37.6 | 43.8 |
| Released | 2024-09-06 | 2026-06-01 |
| Weights | Open | Open |
| Context window | — | 1M |
| Max output | — | 512K |
| Input $ / M tokens | — | $0.30 |
| Output $ / M tokens | — | $1.20 |
| Results tracked | 22 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 31.7 (#281), MiniMax-M3: 41.8 (#118)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Coding | 1309 | 1469 |
| FrontierCode | — | 14.7% |
| Aider Polyglot | 17.8% | — |
| LMArena WebDev | — | 1482 |
| SciCode | — | 47.1% |
| BigCodeBench Instruct | 48.6% | — |
| BigCodeBench Complete | 53.2% | — |
| ALE-Bench | — | 640.02 |
| HumanEval+ | 83.5% | — |
| MBPP+ | 74.1% | — |
Agentic & Tool Use Not comparable
DeepSeek-V2.5 (Sep 2024): —, MiniMax-M3: 22.6 (#130)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| GBAEval | — | 0.9% |
| Vending-Bench 2 | — | 2,158 |
Reasoning MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 25.6 (#145), MiniMax-M3: 30.1 (#87)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Hard Prompts | 1289 | 1447 |
| SimpleBench | — | 45.8% |
| NYT Connections (extended) | — | 65.1% |
| CritPt | — | 3.7% |
| Chess Puzzles | — | 14% |
| Mystery Game Puzzles | — | 8% |
| DTBench | — | 78.9% |
| LMCA | — | 33.7% |
| Surface Evolver Bench | — | 55% |
| Epoch Capabilities Index | — | 146.95 |
| ForecastBench | — | 61.4 |
Math MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 35.9 (#177), MiniMax-M3: 40.0 (#95)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Math | 1288 | 1429 |
| OTIS Mock AIME 2024-2025 | — | 71.1% |
| ProofBench | — | 18% |
Knowledge MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 34.8 (#193), MiniMax-M3: 58.4 (#35)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Expert | 1266 | 1461 |
| GPQA Diamond | — | 90.9% |
Multimodal Not comparable
DeepSeek-V2.5 (Sep 2024): —, MiniMax-M3: 40.2 (#51)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1253 |
| LMArena Document | — | 1435 |
Multilingual MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 42.5 (#193), MiniMax-M3: 53.0 (#75)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1273 | 1420 |
| LMArena Chinese | 1318 | 1463 |
| LMArena French | 1289 | 1447 |
| LMArena German | 1258 | 1426 |
| LMArena Japanese | 1228 | 1381 |
| LMArena Korean | 1209 | 1372 |
| LMArena Russian | 1289 | 1428 |
| LMArena Spanish | 1248 | 1432 |
Instruction Following MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 67.5 (#194), MiniMax-M3: 75.5 (#62)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1280 | 1433 |
Long Context MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 39.5 (#174), MiniMax-M3: 44.2 (#72)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1301 | 1445 |
Writing & Preference MiniMax-M3 leads
DeepSeek-V2.5 (Sep 2024): 49.8 (#187), MiniMax-M3: 62.1 (#83)
| Benchmark | DeepSeek-V2.5 (Sep 2024) | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1294 | 1433 |
| LMArena Creative Writing | 1285 | 1404 |
| LMArena Multi-Turn | 1297 | 1442 |
| EQ-Bench 4 | — | 1150 |
Frequently asked questions
Is DeepSeek-V2.5 (Sep 2024) better than MiniMax-M3?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 37.6 on the Noometry Index.
Is DeepSeek-V2.5 (Sep 2024) or MiniMax-M3 better for coding?
MiniMax-M3 scores higher on coding benchmarks: 41.8 versus 31.7 in the Noometry coding category.
How many benchmarks do DeepSeek-V2.5 (Sep 2024) and MiniMax-M3 share?
17 benchmarks have published results for both models. DeepSeek-V2.5 (Sep 2024) has 22 scored results on Noometry and MiniMax-M3 has 41.