Model comparison
DeepSeek-V3 vs MiniMax-M2.7
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 37.7 on the Noometry Index.
Last verified . 22 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 22 benchmarks with published results for both. DeepSeek-V3 scores higher in 3 categories and MiniMax-M2.7 in 5 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in long context, where MiniMax-M2.7 leads 43.3 to 34.0.
- The biggest single-benchmark swing is SciCode: 35.8% for DeepSeek-V3 and 47% for MiniMax-M2.7.
- DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.7.
- MiniMax-M2.7 accepts more context: 205K tokens versus 164K.
Side by side
| DeepSeek-V3 | MiniMax-M2.7 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 39.5 | 37.7 |
| Released | 2024-12-26 | 2026-03-18 |
| Weights | Open | Open |
| Context window | 164K | 205K |
| Max output | 164K | 131K |
| Input $ / M tokens | $0.24 | $0.30 |
| Output $ / M tokens | $0.90 | $1.20 |
| Results tracked | 60 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek-V3: 42.3 (#106), MiniMax-M2.7: 41.8 (#120)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| SciCode | 35.8% | 47% |
| WeirdML | 36.1% | 37% |
| LMArena Coding | 1368 | 1454 |
| Aider Polyglot | 55.1% | — |
| LMArena WebDev | — | 1398 |
| BigCodeBench Instruct | 50% | — |
| LiveBench Coding | 70.9% | — |
| BigCodeBench Complete | 62.2% | — |
| ALE-Bench | — | 599.25 |
| HumanEval+ | 86.6% | — |
| MBPP+ | 73% | — |
Agentic & Tool Use Not comparable
DeepSeek-V3: —, MiniMax-M2.7: 25.1 (#111)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| Terminal-Bench | — | 45.1% |
| ExploitBench | — | 13.3% |
| GBAEval | — | 0% |
| METR Time Horizons | 49.6% | — |
Reasoning Too close to call
DeepSeek-V3: 20.5 (#236), MiniMax-M2.7: 19.7 (#253)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| CritPt | 0% | 0.6% |
| LMArena Hard Prompts | 1365 | 1422 |
| Epoch Capabilities Index | 135.94 | 145.85 |
| SimpleBench | 27.2% | — |
| Kagi LLM Benchmark | 52.3% | — |
| NYT Connections (extended) | — | 24.7% |
| Thematic Generalization | — | 39.3% |
| LiveBench Reasoning | 65.8% | — |
| DTBench | 64.8% | — |
| LiveBench Data Analysis | 60.9% | — |
| LMCA | 15.5% | — |
| BIG-Bench Hard | 87.5% | — |
| ForecastBench | 59.1 | — |
| HellaSwag | 88.9% | — |
| LiveBench | 66.9% | — |
| PIQA | 84.7% | — |
| WinoGrande | 85.2% | — |
Math DeepSeek-V3 leads
DeepSeek-V3: 32.1 (#219), MiniMax-M2.7: 25.9 (#263)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| LMArena Math | 1373 | 1420 |
| OTIS Mock AIME 2024-2025 | 37.8% | — |
| ProofBench | — | 3% |
| Omni-MATH | 40.3% | — |
| LiveBench Math | 73.5% | — |
| MATH Level 5 | 75.5% | — |
| FrontierMath (Feb 2025 set) | 1.7% | — |
Knowledge Too close to call
DeepSeek-V3: 37.5 (#155), MiniMax-M2.7: 37.7 (#152)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| Vectara Hallucination Rate | 6.1% | 12.9% |
| LMArena Expert | 1351 | 1444 |
| GPQA Diamond | 67.6% | — |
| MMLU-Pro | 72.3% | — |
| Confabulations | 26.1% | — |
| GPQA (HELM) | 53.8% | — |
| ARC (AI2) Challenge | 95.3% | — |
| MMLU | 87.2% | — |
| TriviaQA | 82.9% | — |
Multilingual MiniMax-M2.7 leads
DeepSeek-V3: 48.5 (#143), MiniMax-M2.7: 50.3 (#123)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| LMArena Non-English | 1358 | 1382 |
| LMArena Chinese | 1391 | 1441 |
| LMArena French | 1385 | 1421 |
| LMArena German | 1374 | 1398 |
| LMArena Japanese | 1333 | 1262 |
| LMArena Korean | 1319 | 1313 |
| LMArena Russian | 1373 | 1383 |
| LMArena Spanish | 1358 | 1403 |
Instruction Following MiniMax-M2.7 leads
DeepSeek-V3: 72.8 (#130), MiniMax-M2.7: 74.1 (#103)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| LMArena Instruction Following | 1345 | 1405 |
| LiveBench Instruction Following | 81.5% | — |
| IFEval | 83.2% | — |
Long Context MiniMax-M2.7 leads
DeepSeek-V3: 34.0 (#253), MiniMax-M2.7: 43.3 (#99)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| LMArena Longer Query | 1352 | 1419 |
| Fiction.LiveBench | 50% | — |
Writing & Preference MiniMax-M2.7 leads
DeepSeek-V3: 57.4 (#130), MiniMax-M2.7: 58.9 (#112)
| Benchmark | DeepSeek-V3 | MiniMax-M2.7 |
|---|---|---|
| LMArena Text | 1375 | 1405 |
| LMArena Creative Writing | 1364 | 1354 |
| LMArena Multi-Turn | 1389 | 1412 |
| Short-Story Creative Writing | 77% | — |
| EQ-Bench Creative Writing | 1472 | — |
| WildBench | 83% | — |
| LiveBench Language | 49.1% | — |
Frequently asked questions
Is DeepSeek-V3 better than MiniMax-M2.7?
DeepSeek-V3 is the stronger model overall, scoring 39.5 to 37.7 on the Noometry Index.
Which is cheaper, DeepSeek-V3 or MiniMax-M2.7?
DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; MiniMax-M2.7 lists at $0.30 and $1.20.
Is DeepSeek-V3 or MiniMax-M2.7 better for coding?
They score almost the same on coding (42.3 vs 41.8); test both on your own repository before choosing.
Which has the bigger context window?
MiniMax-M2.7 does, with 205K tokens against 164K.
How many benchmarks do DeepSeek-V3 and MiniMax-M2.7 share?
22 benchmarks have published results for both models. DeepSeek-V3 has 60 scored results on Noometry and MiniMax-M2.7 has 30.