All tasks / Comparisons

MiniMax-M3 vs Qwen3.7-Plus

Compared only on public benchmark results, across the 6 tasks where both are listed: Qwen3.7-Plus is a tier higher in 1 of 6 shared tasks, they share a tier in 5.

MiniMax-M3

MiniMax's open-weights MoE model line; M3 is natively multimodal with ~428B total and ~23B activated parameters.

Vendor: MiniMax
Weights: Open: run it yourself
Released: 1 Jun 2026
API price: $0.3 in / $1.2 out per 1M tokens
GPU memory to run locally: ~245 GB
License: MiniMax Community License
Last verified: 18 Oct 2026

Alternatives to MiniMax-M3

Qwen3.7-Plus

Alibaba's closed mid-tier Qwen API model on Model Studio, below Qwen-Max in price.

Vendor: Alibaba (Qwen)
Weights: Closed: API and apps only
API price: $0.32 in / $1.28 out per 1M tokens (OpenRouter listed price; official Model Studio price not confirmed)
Last verified: 19 Oct 2026

Alternatives to Qwen3.7-Plus

Task by task

Task	MiniMax-M3	Qwen3.7-Plus	Higher tier
Chat & assistants Models	A1,441 on LMArena Text Arena (overall, style control) · Local	A1,456 on LMArena Text Arena (overall, style control) · Cloud	Same tier
Writing Models	B1,405 on LMArena Text Arena: Creative Writing (style control) · Local	A1,432 on LMArena Text Arena: Creative Writing (style control) · Cloud	Qwen3.7-Plus
Coding Models	C1,482 on LMArena WebDev / Code Arena (overall) · Local	C25 on Artificial Analysis Intelligence Index · Cloud	Same tier
Image understanding Models	A1,236 on LMArena Vision Arena (overall, style control) · Local	A1,264 on LMArena Vision Arena (overall, style control) · Cloud	Same tier
Vector graphics & logos Models	C1,160 on Design Arena SVG · Local	C1,208 on Design Arena SVG · Cloud	Same tier
Browser & computer agents Models	C4.6% on OSWorld 2.0 (end-to-end models) · Local	C2.8% on OSWorld 2.0 (end-to-end models) · Cloud	Same tier

A tier spans a range of scores, and small gaps on a leaderboard are often within its margin of error, so only a different tier counts as ahead. How we rank.

Same benchmark, both results

Benchmark	MiniMax-M3	Qwen3.7-Plus	Snapshot
Artificial Analysis Intelligence Index Composite index over Artificial Analysis' own evaluation suite, higher is better	29 MiniMax-M3	25 Qwen3.7 Plus	19 Oct 2026
Artificial Analysis MMMU-Pro (independent evaluation) Accuracy on MMMU-Pro (multi-discipline multimodal questions over images), as run by Artificial Analysis, higher is better	78.6% MiniMax-M3	80.5% Qwen3.7 Plus	19 Oct 2026
Design Arena SVG Elo rating from crowdsourced blind pairwise votes on SVG markup that models generate from a prompt (icons, illustrations, diagrams, artistic compositions), higher is better	1,160 MiniMax M3	1,208 Qwen3.7 Plus	19 Oct 2026
LMArena Text Arena (overall, style control) Arena score (Bradley-Terry rating from pairwise human votes, style-controlled default view), higher is better	1,441 minimax-m3	1,456 qwen3.7-plus	15 Oct 2026
LMArena Text Arena: Creative Writing (style control) Arena score (Bradley-Terry rating from pairwise human votes, creative-writing prompts only, style-controlled view), higher is better	1,405 minimax-m3	1,432 qwen3.7-plus	15 Oct 2026
LMArena Vision Arena (overall, style control) Arena score (Bradley-Terry rating from pairwise human votes on image + text prompts, style-controlled view), higher is better	1,236 minimax-m3	1,264 qwen3.7-plus	9 Oct 2026
OSWorld 2.0 (end-to-end models) Binary accuracy: share of 108 long-horizon computer-use workflows fully completed (default view: 500-step budget, full task set), higher is better	4.6% MiniMax M3	2.8% Qwen 3.7-Plus	15 Oct 2026