Model comparison
DeepSeek-R1 vs MiniMax-M2.7
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 37.4 on the Noometry Index. MiniMax-M2.7 costs 1.7× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 23 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 23 benchmarks with published results for both. DeepSeek-R1 scores higher in 7 categories and MiniMax-M2.7 in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-R1 leads 42.0 to 24.9.
- The biggest single-benchmark swing is SciCode: 35.7% for DeepSeek-R1 and 47% for MiniMax-M2.7.
- MiniMax-M2.7 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- MiniMax-M2.7 accepts more context: 205K tokens versus 164K.
- MiniMax-M2.7 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | MiniMax-M2.7 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 42.0 | 37.4 |
| Released | 2025-01-20 | 2026-03-18 |
| Weights | Proprietary | Open |
| Context window | 164K | 205K |
| Max output | 64K | 131K |
| Input $ / M tokens | $0.50 | $0.30 |
| Output $ / M tokens | $2.15 | $1.20 |
| Results tracked | 52 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 45.9 (#63), MiniMax-M2.7: 41.4 (#115)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| SciCode | 35.7% | 47% |
| WeirdML | 41.6% | 37% |
| LMArena Coding | 1427 | 1454 |
| ALE-Bench | 804.12 | 599.25 |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1397 |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.4 (#68), MiniMax-M2.7: 24.9 (#103)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| Terminal-Bench | — | 45.1% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| ExploitBench | — | 13.3% |
| GBAEval | — | 0% |
| METR Time Horizons | 53.8% | — |
Reasoning Too close to call
DeepSeek-R1: 18.7 (#253), MiniMax-M2.7: 19.0 (#248)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| CritPt | 1.1% | 0.6% |
| LMArena Hard Prompts | 1415 | 1422 |
| Epoch Capabilities Index | 141.3 | 145.84 |
| ARC-AGI-2 | 1.3% | — |
| SimpleBench | 40.8% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 24.7% |
| ARC-AGI-1 | 21.2% | — |
| Thematic Generalization | — | 39.3% |
| LiveBench Reasoning | 83.2% | — |
| LiveBench Data Analysis | 69.8% | — |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 42.0 (#70), MiniMax-M2.7: 24.9 (#254)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| LMArena Math | 1401 | 1420 |
| OTIS Mock AIME 2024-2025 | 66.4% | — |
| ProofBench | — | 3% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.3 (#81), MiniMax-M2.7: 37.1 (#151)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| Vectara Hallucination Rate | 11.3% | 12.9% |
| LMArena Expert | 1394 | 1443 |
| GPQA Diamond | 76.3% | — |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| GPQA (HELM) | 66.6% | — |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.6 (#80), MiniMax-M2.7: 50.5 (#120)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| LMArena Non-English | 1412 | 1382 |
| LMArena Chinese | 1443 | 1439 |
| LMArena French | 1415 | 1422 |
| LMArena German | 1405 | 1398 |
| LMArena Japanese | 1394 | 1261 |
| LMArena Korean | 1359 | 1313 |
| LMArena Russian | 1423 | 1383 |
| LMArena Spanish | 1410 | 1403 |
Instruction Following MiniMax-M2.7 leads
DeepSeek-R1: 72.2 (#138), MiniMax-M2.7: 74.2 (#101)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| LMArena Instruction Following | 1382 | 1404 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), MiniMax-M2.7: 43.3 (#94)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| LMArena Longer Query | 1391 | 1419 |
| Fiction.LiveBench | 75% | — |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.5 (#84), MiniMax-M2.7: 59.1 (#107)
| Benchmark | DeepSeek-R1 | MiniMax-M2.7 |
|---|---|---|
| LMArena Text | 1428 | 1404 |
| LMArena Creative Writing | 1406 | 1354 |
| LMArena Multi-Turn | 1404 | 1413 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than MiniMax-M2.7?
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 37.4 on the Noometry Index. MiniMax-M2.7 costs 1.7× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 or MiniMax-M2.7?
MiniMax-M2.7 is cheaper. It lists at $0.30 per million input tokens and $1.20 per million output tokens; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or MiniMax-M2.7 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 45.9 versus 41.4 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.7 does, with 205K tokens against 164K.
How many benchmarks do DeepSeek-R1 and MiniMax-M2.7 share?
23 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and MiniMax-M2.7 has 30.