Model comparison
DeepSeek V4 Flash vs MiniMax-M2.5
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 38.3 on the Noometry Index.
Last verified . 26 shared benchmarks.
DeepSeek
MiniMaxSummary
- They share 26 benchmarks with published results for both. DeepSeek V4 Flash scores higher in 7 categories and MiniMax-M2.5 in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Flash leads 53.7 to 17.5.
- The biggest single-benchmark swing is NYT Connections (extended): 89.6% for DeepSeek V4 Flash and 16.8% for MiniMax-M2.5.
- DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.30 / $1.20 for MiniMax-M2.5.
- DeepSeek V4 Flash accepts more context: 1M tokens versus 205K.
Side by side
| DeepSeek V4 Flash | MiniMax-M2.5 | |
|---|---|---|
| Provider | DeepSeek | MiniMax |
| Noometry Index | 53.6 | 38.3 |
| Released | 2026-04-24 | 2026-02-12 |
| Weights | Open | Open |
| Context window | 1M | 205K |
| Max output | 393K | 131K |
| Input $ / M tokens | $0.15 | $0.30 |
| Output $ / M tokens | $0.60 | $1.20 |
| Results tracked | 41 | 33 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek V4 Flash: 47.9 (#59), MiniMax-M2.5: 48.1 (#58)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena WebDev | 1582 | 1387 |
| LMArena Coding | 1457 | 1381 |
| ALE-Bench | 1,306 | 618.17 |
| FrontierCode | 18.8% | — |
| SWE-bench Verified (bash only) | — | 75.8% |
| SWE-bench Multilingual | — | 68.3% |
| SciCode | 49.9% | — |
| WeirdML | 63% | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Flash: —, MiniMax-M2.5: 30.4 (#77)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| Terminal-Bench | — | 42.7% |
| Vending-Bench 2 | — | -23.16 |
Reasoning DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.7 (#30), MiniMax-M2.5: 17.5 (#292)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| ARC-AGI-2 | 61.4% | 4.9% |
| Kagi LLM Benchmark | 52.2% | 55.2% |
| NYT Connections (extended) | 89.6% | 16.8% |
| ARC-AGI-1 | 89% | 63.7% |
| LMArena Hard Prompts | 1444 | 1372 |
| Epoch Capabilities Index | 154.49 | 146.68 |
| SimpleBench | 61.1% | — |
| CritPt | 16.6% | — |
| Chess Puzzles | 33% | — |
| Mystery Game Puzzles | 34% | — |
| DTBench | 90.9% | — |
| LMCA | 41.7% | — |
Math DeepSeek V4 Flash leads
DeepSeek V4 Flash: 60.3 (#37), MiniMax-M2.5: 26.9 (#253)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| ProofBench | 56% | 4% |
| LMArena Math | 1427 | 1378 |
| FrontierMath (Tiers 1-3) | 57.5% | — |
| FrontierMath Tier 4 | 24.4% | — |
| MathArena Final-Answer Competitions | 76.5% | — |
| OTIS Mock AIME 2024-2025 | 94.4% | — |
Knowledge DeepSeek V4 Flash leads
DeepSeek V4 Flash: 55.4 (#48), MiniMax-M2.5: 39.2 (#135)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena Expert | 1441 | 1379 |
| GPQA Diamond | 91% | — |
| SimpleQA Verified | 33.6% | — |
| Vectara Hallucination Rate | — | 9.1% |
Multilingual DeepSeek V4 Flash leads
DeepSeek V4 Flash: 53.0 (#72), MiniMax-M2.5: 47.1 (#152)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena Non-English | 1420 | 1338 |
| LMArena Chinese | 1468 | 1393 |
| LMArena French | 1439 | 1362 |
| LMArena German | 1418 | 1362 |
| LMArena Japanese | 1406 | 1171 |
| LMArena Korean | 1384 | 1232 |
| LMArena Russian | 1428 | 1358 |
| LMArena Spanish | 1436 | 1354 |
Instruction Following DeepSeek V4 Flash leads
DeepSeek V4 Flash: 74.9 (#81), MiniMax-M2.5: 71.5 (#148)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena Instruction Following | 1421 | 1353 |
Long Context DeepSeek V4 Flash leads
DeepSeek V4 Flash: 43.8 (#85), MiniMax-M2.5: 37.5 (#216)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena Longer Query | 1434 | 1366 |
| CL-bench | — | 11.4% |
| CL-bench Life | — | 6.3% |
Writing & Preference DeepSeek V4 Flash leads
DeepSeek V4 Flash: 63.8 (#61), MiniMax-M2.5: 53.9 (#153)
| Benchmark | DeepSeek V4 Flash | MiniMax-M2.5 |
|---|---|---|
| LMArena Text | 1432 | 1359 |
| LMArena Creative Writing | 1403 | 1331 |
| EQ-Bench Creative Writing | 1559 | 1361 |
| LMArena Multi-Turn | 1449 | 1364 |
Frequently asked questions
Is DeepSeek V4 Flash better than MiniMax-M2.5?
DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 38.3 on the Noometry Index.
Which is cheaper, DeepSeek V4 Flash or MiniMax-M2.5?
DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; MiniMax-M2.5 lists at $0.30 and $1.20.
Is DeepSeek V4 Flash or MiniMax-M2.5 better for coding?
They score almost the same on coding (47.9 vs 48.1); test both on your own repository before choosing.
Which has the bigger context window?
DeepSeek V4 Flash does, with 1M tokens against 205K.
How many benchmarks do DeepSeek V4 Flash and MiniMax-M2.5 share?
26 benchmarks have published results for both models. DeepSeek V4 Flash has 41 scored results on Noometry and MiniMax-M2.5 has 33.