Model comparison
DeepSeek-V3.2-Exp vs MiniMax-M2
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 37.4 on the Noometry Index.
Last verified . 21 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 21 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 9 categories and MiniMax-M2 in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where DeepSeek-V3.2-Exp leads 51.7 to 37.0.
- The biggest single-benchmark swing is NYT Connections (extended): 36.7% for DeepSeek-V3.2-Exp and 14.8% for MiniMax-M2.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.
- MiniMax-M2 accepts more context: 205K tokens versus 164K.
Side by side
| DeepSeek-V3.2-Exp | MiniMax-M2 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 44.3 | 37.4 |
| Released | 2025-09-29 | 2025-10-27 |
| Weights | Open | Open |
| Context window | 164K | 205K |
| Max output | 66K | 131K |
| Input $ / M tokens | $0.26 | $0.30 |
| Output $ / M tokens | $0.38 | $1.20 |
| Results tracked | 49 | 21 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 46.5 (#65), MiniMax-M2: 39.3 (#159)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| SWE-bench Verified (bash only) | 70% | 61% |
| LMArena WebDev | 1362 | 1297 |
| LMArena Coding | 1454 | 1370 |
| Aider Polyglot | 74.2% | — |
| SWE-bench Multilingual | 59% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 32.7 (#59), MiniMax-M2: 25.1 (#109)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| Terminal-Bench | 39.6% | 30% |
| Vending-Bench 2 | 1,034 | 160.6 |
| APEX-Agents | 21.3% | — |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
Reasoning DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 22.1 (#208), MiniMax-M2: 19.4 (#258)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| Kagi LLM Benchmark | 52.2% | 57.8% |
| NYT Connections (extended) | 36.7% | 14.8% |
| LMArena Hard Prompts | 1434 | 1357 |
| ARC-AGI-2 | 4% | — |
| ARC-AGI-1 | 57% | — |
| CritPt | 2.9% | — |
| Chess Puzzles | 14% | — |
| Thematic Generalization | 65% | — |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
| Epoch Capabilities Index | 146.27 | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 41.7 (#87), MiniMax-M2: 37.3 (#160)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Math | 1435 | 1352 |
| MathArena Final-Answer Competitions | 57.7% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| ProofBench | 8% | — |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.7 (#66), MiniMax-M2: 37.0 (#163)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Expert | 1436 | 1337 |
| GPQA Diamond | 83.4% | — |
| Vectara Hallucination Rate | 5.3% | — |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.2 (#90), MiniMax-M2: 45.3 (#171)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Non-English | 1409 | 1313 |
| LMArena Chinese | 1461 | 1366 |
| LMArena French | 1433 | 1335 |
| LMArena German | 1440 | 1355 |
| LMArena Russian | 1424 | 1331 |
| LMArena Spanish | 1440 | 1326 |
| LMArena Japanese | 1374 | — |
| LMArena Korean | 1371 | — |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.5 (#93), MiniMax-M2: 70.2 (#166)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Instruction Following | 1413 | 1328 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#16), MiniMax-M2: 40.5 (#153)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Longer Query | 1428 | 1331 |
| Fiction.LiveBench | 83.3% | — |
| CL-bench | 13.2% | — |
| CL-bench Life | 9.5% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.4 (#77), MiniMax-M2: 53.0 (#162)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2 |
|---|---|---|
| LMArena Text | 1425 | 1340 |
| LMArena Creative Writing | 1403 | 1286 |
| LMArena Multi-Turn | 1427 | 1361 |
| EQ-Bench Creative Writing | 1515 | — |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than MiniMax-M2?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 37.4 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or MiniMax-M2?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; MiniMax-M2 lists at $0.30 and $1.20.
Is DeepSeek-V3.2-Exp or MiniMax-M2 better for coding?
DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 39.3 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2 does, with 205K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and MiniMax-M2 share?
21 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and MiniMax-M2 has 21.