Model comparison
DeepSeek-V3.2-Exp vs MiniMax-M2.5
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 38.1 on the Noometry Index.
Last verified . 32 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 32 benchmarks with published results for both. DeepSeek-V3.2-Exp scores higher in 8 categories and MiniMax-M2.5 in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where DeepSeek-V3.2-Exp leads 40.8 to 25.8.
- The biggest single-benchmark swing is NYT Connections (extended): 36.7% for DeepSeek-V3.2-Exp and 16.8% for MiniMax-M2.5.
- DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.5.
- MiniMax-M2.5 accepts more context: 205K tokens versus 164K.
Side by side
| DeepSeek-V3.2-Exp | MiniMax-M2.5 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 44.9 | 38.1 |
| Released | 2025-09-29 | 2026-02-12 |
| 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 | 48 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding MiniMax-M2.5 leads
DeepSeek-V3.2-Exp: 46.4 (#61), MiniMax-M2.5: 47.9 (#53)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| SWE-bench Verified (bash only) | 70% | 75.8% |
| LMArena WebDev | 1362 | 1386 |
| SWE-bench Multilingual | 59% | 68.3% |
| LMArena Coding | 1454 | 1380 |
| Aider Polyglot | 74.2% | — |
| SciCode | 38.9% | — |
| WeirdML | 39.5% | — |
| ALE-Bench | — | 618.17 |
Agentic & Tool Use DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 37.9 (#31), MiniMax-M2.5: 30.2 (#70)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | 39.6% | 42.7% |
| Vending-Bench 2 | 1,034 | -23.16 |
| Berkeley Function Calling Leaderboard | 56.7% | — |
| TheAgentCompany | 42.9% | — |
Reasoning DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 22.0 (#185), MiniMax-M2.5: 17.6 (#281)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| ARC-AGI-2 | 4% | 4.9% |
| Kagi LLM Benchmark | 52.2% | 55.2% |
| NYT Connections (extended) | 36.7% | 16.8% |
| ARC-AGI-1 | 57% | 63.7% |
| LMArena Hard Prompts | 1433 | 1372 |
| Epoch Capabilities Index | 146.28 | 146.68 |
| CritPt | 2.9% | — |
| Chess Puzzles | 14% | — |
| Thematic Generalization | 65% | — |
| DTBench | 87.7% | — |
| LMCA | 29.1% | — |
Math DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 40.8 (#75), MiniMax-M2.5: 25.8 (#242)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| ProofBench | 8% | 4% |
| LMArena Math | 1435 | 1378 |
| MathArena Final-Answer Competitions | 57.7% | — |
| OTIS Mock AIME 2024-2025 | 87.8% | — |
| FrontierMath (Feb 2025 set) | 22.1% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 51.6 (#61), MiniMax-M2.5: 38.6 (#132)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| Vectara Hallucination Rate | 5.3% | 9.1% |
| LMArena Expert | 1436 | 1378 |
| GPQA Diamond | 83.4% | — |
Multilingual DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 52.4 (#85), MiniMax-M2.5: 47.3 (#147)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1409 | 1338 |
| LMArena Chinese | 1461 | 1392 |
| LMArena French | 1435 | 1363 |
| LMArena German | 1440 | 1361 |
| LMArena Japanese | 1377 | 1169 |
| LMArena Korean | 1372 | 1232 |
| LMArena Russian | 1423 | 1359 |
| LMArena Spanish | 1441 | 1354 |
Instruction Following DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 74.7 (#89), MiniMax-M2.5: 71.7 (#143)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1413 | 1353 |
Long Context DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 47.6 (#15), MiniMax-M2.5: 37.4 (#211)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| CL-bench | 13.2% | 11.4% |
| CL-bench Life | 9.5% | 6.3% |
| LMArena Longer Query | 1428 | 1367 |
| Fiction.LiveBench | 83.3% | — |
Writing & Preference DeepSeek-V3.2-Exp leads
DeepSeek-V3.2-Exp: 62.5 (#73), MiniMax-M2.5: 54.1 (#149)
| Benchmark | DeepSeek-V3.2-Exp | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1425 | 1359 |
| LMArena Creative Writing | 1403 | 1331 |
| EQ-Bench Creative Writing | 1515 | 1361 |
| LMArena Multi-Turn | 1427 | 1364 |
Frequently asked questions
Is DeepSeek-V3.2-Exp better than MiniMax-M2.5?
DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.9 to 38.1 on the Noometry Index.
Which is cheaper, DeepSeek-V3.2-Exp or MiniMax-M2.5?
DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; MiniMax-M2.5 lists at $0.30 and $1.20.
Is DeepSeek-V3.2-Exp or MiniMax-M2.5 better for coding?
MiniMax-M2.5 scores higher on coding benchmarks: 47.9 versus 46.4 in the Noometry coding category.
Which has the bigger context window?
MiniMax-M2.5 does, with 205K tokens against 164K.
How many benchmarks do DeepSeek-V3.2-Exp and MiniMax-M2.5 share?
32 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 48 scored results on Noometry and MiniMax-M2.5 has 33.