MiniMax M2.5 vs Qwen3.7 Plus
MiniMax M2.5 leads Qwen3.7 Plus on coding, 37.2 to 33.2 on the Vibe Coding Index, at about the same output price ($1.20 vs $1.28 per million). Updated October 22, 2026; 44 benchmarks have a result for both.
MiniMax M2.5 or Qwen3.7 Plus: which to pick
Pick MiniMax M2.5 for coding: it scores 37.2 on the Vibe Coding Index against 33.2 for Qwen3.7 Plus.
Pick Qwen3.7 Plus for a larger context window (1M against 205K tokens) and work measured by Vibe Code Bench (Vals AI) or Terminal-Bench Hard (Artificial Analysis run), where it leads.
- Shared results
- 44
- MiniMax M2.5 only
- 24
- Qwen3.7 Plus only
- 31
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
MiniMax M2.5
MiniMax M2.5 scores 37.2 against 33.2.
3 and 2 benchmarks behind the scores
Intelligence
AA Intelligence Index
Qwen3.7 Plus
Qwen3.7 Plus leads, 25.2 against 22.8.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Qwen3.7 Plus has a AA Agentic Index result, 17.5.
Directional only
Reasoning
GPQA Diamond
No call
Only MiniMax M2.5 has a GPQA Diamond result, 82.1%.
Directional only
Math
AIME
No call
Only MiniMax M2.5 has a AIME result, 88.8%.
Directional only
Knowledge
MMLU Pro
No call
Only MiniMax M2.5 has a MMLU Pro result, 80.1%.
Directional only
Web design
Design Arena: websites
Qwen3.7 Plus
Qwen3.7 Plus leads, 1269 against 1222.
Same board, both measured
Long documents
Listed context window
Qwen3.7 Plus
Qwen3.7 Plus takes 1M tokens against 205K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
MiniMax M2.5
MiniMax M2.5 $0.0009, Qwen3.7 Plus $0.0010 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only MiniMax M2.5 has the number, $0.62.
Directional only
At a glance
- Coding index
- MiniMax M2.5 37.2 from 3 benchmarks, Qwen3.7 Plus 33.2 from 2 benchmarks.
- Widest category
- Coding, 18.5 points apart on average over 4 shared benchmarks.
- Widest single gap
- LMArena Text: Japanese: MiniMax M2.5 1171 against Qwen3.7 Plus 1412, 40.1 points apart on a 0 to 100 scale.
- Output price
- MiniMax M2.5 $1.20 against Qwen3.7 Plus $1.28 per million output tokens, 1.1x apart.
- Context window
- MiniMax M2.5 takes 205K tokens, Qwen3.7 Plus 1M.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | MiniMax M2.5 | Bars | Qwen3.7 Plus | Ties |
|---|---|---|---|---|
| Coding | 0 | 4 | ||
| Intelligence | 0 | 2 | ||
| Agentic work | 1 | 0 | ||
| Reasoning | 0 | 5 | ||
| Math | 0 | 2 | ||
| Knowledge | 0 | 4 | ||
| Web design | 0 | 7 | ||
| General | 1 | 18 | ||
| All shared boards | 2 | 42 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index04
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| SWE-bench Verified | 75.8% | Not measured | Only MiniMax M2.5 measured |
| SWE-bench Multilingual | 68.3% | Not measured | Only MiniMax M2.5 measured |
| LMArena WebDev | 1386 | Not measured | Only MiniMax M2.5 measured |
| AA Coding Index | Not measured | 55.9 | Only Qwen3.7 Plus measured |
| LiveCodeBench (Vals AI) | 79.2% | Not measured | Only MiniMax M2.5 measured |
| SWE-bench Verified (Vals AI) | 74.2% | Not measured | Only MiniMax M2.5 measured |
| Vibe Code Bench (Vals AI) | 14.8% | 46.4% | Qwen3.7 Plus |
| Code migration (Vals AI) | Not measured | 12.9% | Only Qwen3.7 Plus measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 52.8% | Only Qwen3.7 Plus measured |
| SkillsBench (Vals AI) | Not measured | 54.3% | Only Qwen3.7 Plus measured |
| SciCode | Not measured | 46.1% | Only Qwen3.7 Plus measured |
| Terminal-Bench Hard (Artificial Analysis run) | 34.8% | 47.0% | Qwen3.7 Plus |
| Terminal-Bench | 42.7% | Not measured | Only MiniMax M2.5 measured |
| FrontierCode | Not measured | 10.2% | Only Qwen3.7 Plus measured |
| ALE-Bench | 618.2 | Not measured | Only MiniMax M2.5 measured |
| LMArena Text: coding | 1380 | 1473 | Qwen3.7 Plus |
| LMArena Text: software and IT | 1383 | 1473 | Qwen3.7 Plus |
| LMArena WebDev: React | 1374 | Not measured | Only MiniMax M2.5 measured |
| LMArena WebDev: HTML | 1414 | Not measured | Only MiniMax M2.5 measured |
| Neither model has a result on the other 20 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| AA Intelligence Index | 22.8 | 25.2 | Qwen3.7 Plus |
| Epoch Capabilities Index | 146.6 | 147.0 | Qwen3.7 Plus |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 17.5 | Only Qwen3.7 Plus measured |
| Finance Agent (Vals AI) | Not measured | 38.2% | Only Qwen3.7 Plus measured |
| Tax Agent Bench (Vals AI) | Not measured | 10.5% | Only Qwen3.7 Plus measured |
| Legal Research Bench (Vals AI) | Not measured | 16.4% | Only Qwen3.7 Plus measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 49.3% | Only Qwen3.7 Plus measured |
| τ²-Bench Telecom | 95.3% | 93.0% | MiniMax M2.5 |
| OSWorld 2.0 | Not measured | 2.8% | Only Qwen3.7 Plus measured |
| Vending-Bench 2 | -$23 | Not measured | Only MiniMax M2.5 measured |
| Neither model has a result on the other 33 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| GPQA Diamond | 82.1% | Not measured | Only MiniMax M2.5 measured |
| GPQA Diamond (Artificial Analysis run) | 84.8% | 90.0% | Qwen3.7 Plus |
| Humanity's Last Exam (Artificial Analysis run) | 20.5% | 35.6% | Qwen3.7 Plus |
| GPQA Diamond (Epoch AI run) | Not measured | 87.9% | Only Qwen3.7 Plus measured |
| ARC-AGI | 63.7% | Not measured | Only MiniMax M2.5 measured |
| ARC-AGI-2 | 4.9% | Not measured | Only MiniMax M2.5 measured |
| Mystery game puzzles | Not measured | 17.0% | Only Qwen3.7 Plus measured |
| LMArena Text: hard prompts | 1372 | 1460 | Qwen3.7 Plus |
| LMArena Text: expert | 1378 | 1466 | Qwen3.7 Plus |
| LMArena Text: hard prompts in English | 1384 | 1460 | Qwen3.7 Plus |
| LMArena Vision: diagrams | Not measured | 1292 | Only Qwen3.7 Plus measured |
| LMArena Vision: homework | Not measured | 1295 | Only Qwen3.7 Plus measured |
| Kagi LLM Benchmark | 55.2% | Not measured | Only MiniMax M2.5 measured |
| Neither model has a result on the other 9 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| AIME | 88.8% | Not measured | Only MiniMax M2.5 measured |
| OTIS Mock AIME | Not measured | 93.3% | Only Qwen3.7 Plus measured |
| FrontierMath, tiers 1 to 3 | Not measured | 34.4% | Only Qwen3.7 Plus measured |
| ProofBench | 4.0% | Not measured | Only MiniMax M2.5 measured |
| LMArena Text: math | 1378 | 1467 | Qwen3.7 Plus |
| LMArena Text: mathematical work | 1378 | 1470 | Qwen3.7 Plus |
| Neither model has a result on the other 7 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| MMLU Pro | 80.1% | Not measured | Only MiniMax M2.5 measured |
| LegalBench (Vals AI) | 80.0% | Not measured | Only MiniMax M2.5 measured |
| CaseLaw (Vals AI) | 53.5% | Not measured | Only MiniMax M2.5 measured |
| CorpFin (Vals AI) | 59.6% | Not measured | Only MiniMax M2.5 measured |
| TaxEval (Vals AI) | 68.2% | Not measured | Only MiniMax M2.5 measured |
| MortgageTax (Vals AI) | Not measured | 66.2% | Only Qwen3.7 Plus measured |
| MedQA (Vals AI) | 92.5% | Not measured | Only MiniMax M2.5 measured |
| LMArena Text: science | 1369 | 1464 | Qwen3.7 Plus |
| LMArena Text: medicine and healthcare | 1357 | 1464 | Qwen3.7 Plus |
| LMArena Text: legal and government | 1366 | 1451 | Qwen3.7 Plus |
| LMArena Text: business and finance | 1356 | 1435 | Qwen3.7 Plus |
| Neither model has a result on the other 7 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index07
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| Design Arena: websites | 1222 | 1269 | Qwen3.7 Plus |
| Design Arena: UI components | 1173 | 1258 | Qwen3.7 Plus |
| Design Arena: game dev | 1176 | 1261 | Qwen3.7 Plus |
| Design Arena: data viz | 1172 | 1276 | Qwen3.7 Plus |
| Design Arena: 3D | 1170 | 1238 | Qwen3.7 Plus |
| Design Arena: all code categories | 1207 | 1269 | Qwen3.7 Plus |
| Design Arena: SVG | 1150 | 1208 | Qwen3.7 Plus |
| Design Arena: ASCII art | Not measured | 1143 | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index118
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| SAGE (Vals AI) | Not measured | 39.3% | Only Qwen3.7 Plus measured |
| Vals Multimodal Index | Not measured | 53.9% | Only Qwen3.7 Plus measured |
| IFBench | 71.6% | 78.0% | Qwen3.7 Plus |
| Long context reasoning (AA-LCR) | 73.3% | 73.0% | MiniMax M2.5 |
| DTBench | Not measured | 84.0% | Only Qwen3.7 Plus measured |
| LMCA | Not measured | 37.6% | Only Qwen3.7 Plus measured |
| CL-bench | 11.4% | Not measured | Only MiniMax M2.5 measured |
| CL-bench Life | 6.3% | Not measured | Only MiniMax M2.5 measured |
| LMArena Text | 1359 | 1454 | Qwen3.7 Plus |
| LMArena Text: instruction following | 1353 | 1439 | Qwen3.7 Plus |
| LMArena Text: creative writing | 1331 | 1438 | Qwen3.7 Plus |
| LMArena Vision | Not measured | 1278 | Only Qwen3.7 Plus measured |
| LMArena Text: English | 1380 | 1458 | Qwen3.7 Plus |
| LMArena Text: non-English | 1338 | 1445 | Qwen3.7 Plus |
| LMArena Text: Chinese | 1393 | 1509 | Qwen3.7 Plus |
| LMArena Text: German | 1361 | 1471 | Qwen3.7 Plus |
| LMArena Text: Japanese | 1171 | 1412 | Qwen3.7 Plus |
| LMArena Text: Korean | 1233 | 1413 | Qwen3.7 Plus |
| LMArena Text: French | 1363 | 1472 | Qwen3.7 Plus |
| LMArena Text: Spanish | 1353 | 1457 | Qwen3.7 Plus |
| LMArena Text: Russian | 1359 | 1457 | Qwen3.7 Plus |
| LMArena Text: Polish | 1334 | 1449 | Qwen3.7 Plus |
| LMArena Text: multi-turn | 1364 | 1459 | Qwen3.7 Plus |
| LMArena Text: longer queries | 1366 | 1454 | Qwen3.7 Plus |
| LMArena Text: writing and language | 1345 | 1437 | Qwen3.7 Plus |
| LMArena Text: entertainment and media | 1329 | 1422 | Qwen3.7 Plus |
| LMArena Vision: English | Not measured | 1280 | Only Qwen3.7 Plus measured |
| LMArena Vision: Chinese | Not measured | 1332 | Only Qwen3.7 Plus measured |
| LMArena Vision: reading text | Not measured | 1288 | Only Qwen3.7 Plus measured |
| LMArena Vision: humor | Not measured | 1261 | Only Qwen3.7 Plus measured |
| LMArena Vision: creative writing | Not measured | 1258 | Only Qwen3.7 Plus measured |
| LMArena Document | Not measured | 1444 | Only Qwen3.7 Plus measured |
| Neither model has a result on the other 14 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | MiniMax M2.5 | Qwen3.7 Plus | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
MiniMax M2.5 costs less
Repository review
50K input + 3K output tokens
MiniMax M2.5 costs less
Agent loop
200K input + 10K output tokens
MiniMax M2.5 costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | MiniMax M2.5 | Qwen3.7 Plus |
|---|---|---|
| Maker | MiniMax | Qwen |
| Weights | Open weights | No public weights listed |
| Released | February 12, 2026 | June 3, 2026 |
| Context window | 205K | 1M |
| Input $/Mtok | $0.30 | $0.32 |
| Output $/Mtok | $1.20 | $1.28 |
| Coding index | 37.2 | 33.2 |
| Independent measurements behind it | 3 | 2 |
| Cost per solved task | $0.62 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is MiniMax M2.5 better than Qwen3.7 Plus for coding?
Pick MiniMax M2.5 for coding: it scores 37.2 on the Vibe Coding Index against 33.2 for Qwen3.7 Plus. Pick Qwen3.7 Plus for a larger context window (1M against 205K tokens) and work measured by Vibe Code Bench (Vals AI) or Terminal-Bench Hard (Artificial Analysis run), where it leads.
Which is cheaper, MiniMax M2.5 or Qwen3.7 Plus?
MiniMax M2.5 costs $0.30 per million input tokens and $1.20 per million output tokens; Qwen3.7 Plus costs $0.32 and $1.28. Cost per solved task is known for MiniMax M2.5 only.
Which has the bigger context window, MiniMax M2.5 or Qwen3.7 Plus?
Qwen3.7 Plus, with 1M tokens against 205K.
How many benchmarks is the MiniMax M2.5 vs Qwen3.7 Plus comparison based on?
44 benchmarks have a result for both models as of October 22, 2026. MiniMax M2.5's coding index rests on 3 independent benchmarks and Qwen3.7 Plus's on 2. Where only one model has a result, the page says so instead of naming a winner.