MiniMax M2 is 2.7× cheaper per million tokens (blended). Qwen3.8 27B has the larger context window (1,000,000 vs 204,800 tokens, 4.9×).
MiniMax M2 vs Qwen3.8 27B
Live catalog fields. Best-in-row is highlighted.
MiniMax M2open
Minimax · minimax__minimax-m2
Qwen3.8 27Bopen
Qwen · qwen__qwen3.8-27b
Identity
| Field | minimax__minimax-m2 | qwen__qwen3.8-27b |
|---|---|---|
| Provider | Minimax | Qwen |
| Slug | minimax__minimax-m2 | qwen__qwen3.8-27b |
| Status | live | live |
| Open weights | Yes | Yes |
| License | — | — |
| HuggingFace | MiniMaxAI/MiniMax-M2 | Qwen/Qwen3.8-27B |
Quality
| Field | minimax__minimax-m2 | qwen__qwen3.8-27b |
|---|---|---|
| Quality index (0–100) | — | 93.0 |
Cost
| Field | minimax__minimax-m2 | qwen__qwen3.8-27b |
|---|---|---|
| Input $ / 1M tokens | $0.26 | $0.42 |
| Cached input $ / 1M | — | $0.09 |
| Output $ / 1M tokens | $1.02 | $3.00 |
Context
| Field | minimax__minimax-m2 | qwen__qwen3.8-27b |
|---|---|---|
| Max context tokens | 204,800 | 1,000,000 |
| Max output tokens | 131,072 | 131,072 |
Modalities
| Field | minimax__minimax-m2 | qwen__qwen3.8-27b |
|---|---|---|
| Input | text | text, image, video |
| Output | text | text |