Model comparison
DeepSeek-R1 vs MiniMax-M3
MiniMax-M3 is the stronger model overall, scoring 43.8 to 42.3 on the Noometry Index.
Last verified . 25 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 25 benchmarks with published results for both. DeepSeek-R1 scores higher in 4 categories and MiniMax-M3 in 5 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where MiniMax-M3 leads 58.4 to 44.5.
- The biggest single-benchmark swing is GPQA Diamond: 76.3% for DeepSeek-R1 and 90.9% for MiniMax-M3.
- MiniMax-M3 is cheaper at $0.30 / $1.20 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- MiniMax-M3 accepts more context: 1M tokens versus 164K.
- MiniMax-M3 has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | MiniMax-M3 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 42.3 | 43.8 |
| Released | 2025-01-20 | 2026-06-01 |
| Weights | Proprietary | Open |
| Context window | 164K | 1M |
| Max output | 64K | 512K |
| Input $ / M tokens | $0.50 | $0.30 |
| Output $ / M tokens | $2.15 | $1.20 |
| Results tracked | 52 | 41 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 46.3 (#68), MiniMax-M3: 41.8 (#118)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| SciCode | 35.7% | 47.1% |
| LMArena Coding | 1427 | 1469 |
| ALE-Bench | 804.12 | 640.02 |
| FrontierCode | — | 14.7% |
| Aider Polyglot | 71.4% | — |
| LMArena WebDev | — | 1482 |
| WeirdML | 41.6% | — |
| LiveBench Coding | 66.7% | — |
| AlgoTune | 1.7 | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.7 (#75), MiniMax-M3: 22.6 (#130)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| APEX-Agents | — | 37.7% |
| OSWorld 2.0 | — | 4.6% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| GBAEval | — | 0.9% |
| METR Time Horizons | 53.8% | — |
| Vending-Bench 2 | — | 2,158 |
Reasoning MiniMax-M3 leads
DeepSeek-R1: 18.6 (#278), MiniMax-M3: 30.1 (#87)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| SimpleBench | 40.8% | 45.8% |
| CritPt | 1.1% | 3.7% |
| LMArena Hard Prompts | 1416 | 1447 |
| Epoch Capabilities Index | 141.29 | 146.95 |
| ForecastBench | 60 | 61.4 |
| ARC-AGI-2 | 1.3% | — |
| Kagi LLM Benchmark | 69.4% | — |
| NYT Connections (extended) | — | 65.1% |
| ARC-AGI-1 | 21.2% | — |
| Chess Puzzles | — | 14% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 8% |
| DTBench | — | 78.9% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 33.7% |
| Surface Evolver Bench | — | 55% |
| LiveBench | 71.6% | — |
Math DeepSeek-R1 leads
DeepSeek-R1: 43.8 (#79), MiniMax-M3: 40.0 (#95)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 71.1% |
| LMArena Math | 1400 | 1429 |
| ProofBench | — | 18% |
| Omni-MATH | 42.4% | — |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge MiniMax-M3 leads
DeepSeek-R1: 44.5 (#87), MiniMax-M3: 58.4 (#35)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| GPQA Diamond | 76.3% | 90.9% |
| LMArena Expert | 1394 | 1461 |
| MMLU-Pro | 79.3% | — |
| Confabulations | 12.7% | — |
| Vectara Hallucination Rate | 11.3% | — |
| GPQA (HELM) | 66.6% | — |
Multimodal Not comparable
DeepSeek-R1: —, MiniMax-M3: 40.2 (#51)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| LMArena Vision | — | 1253 |
| LMArena Document | — | 1435 |
Multilingual Too close to call
DeepSeek-R1: 52.4 (#85), MiniMax-M3: 53.0 (#75)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| LMArena Non-English | 1412 | 1420 |
| LMArena Chinese | 1442 | 1463 |
| LMArena French | 1417 | 1447 |
| LMArena German | 1404 | 1426 |
| LMArena Japanese | 1391 | 1381 |
| LMArena Korean | 1360 | 1372 |
| LMArena Russian | 1423 | 1428 |
| LMArena Spanish | 1411 | 1432 |
Instruction Following MiniMax-M3 leads
DeepSeek-R1: 72.0 (#143), MiniMax-M3: 75.5 (#62)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| LMArena Instruction Following | 1382 | 1433 |
| LiveBench Instruction Following | 80.5% | — |
| IFEval | 78.4% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), MiniMax-M3: 44.2 (#72)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| LMArena Longer Query | 1391 | 1445 |
| Fiction.LiveBench | 75% | — |
Writing & Preference Too close to call
DeepSeek-R1: 61.4 (#88), MiniMax-M3: 62.1 (#83)
| Benchmark | DeepSeek-R1 | MiniMax-M3 |
|---|---|---|
| LMArena Text | 1428 | 1433 |
| LMArena Creative Writing | 1405 | 1404 |
| LMArena Multi-Turn | 1405 | 1442 |
| Short-Story Creative Writing | 83% | — |
| EQ-Bench Creative Writing | 1500 | — |
| WildBench | 82.8% | — |
| EQ-Bench 4 | — | 1150 |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than MiniMax-M3?
MiniMax-M3 is the stronger model overall, scoring 43.8 to 42.3 on the Noometry Index.
Which is cheaper, DeepSeek-R1 or MiniMax-M3?
MiniMax-M3 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-M3 better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 41.8 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M3 does, with 1M tokens against 164K.
How many benchmarks do DeepSeek-R1 and MiniMax-M3 share?
25 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and MiniMax-M3 has 41.