MiniMax-M2.7 vs GPT-5
API pricing, benchmark indices and specs — updated 19 October 2026.
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | MiniMax-M2.7 | GPT-5 |
|---|---|---|
| Deployment | Open weights · also via API | API-only (closed weights) |
| Parameters | 228.7 B | n/a — API-only |
| Architecture | MiniMaxM2ForCausalLM | n/a — API-only |
| Context length | 204,800 | 400,000 |
| Licence | other | Proprietary |
| Commercial use | Unknown | Per API terms |
| Languages | — | n/a — API-only |
| Downloads 30d | 438,954 | n/a — API-only |
| Downloads all time | 7.9 M | n/a — API-only |
| Quantizations on the Hub | SAFETENSORS | n/a — API-only |
| Gated | No | n/a — API-only |
| First seen on the Hub | 2026-04-09 | n/a — API-only |
| API input price / 1M tokens | $0.30 | $1.25 |
| API output price / 1M tokens | $1.20 | $10.00 |
| Est. cost per 1M input + 1M output tokens | $1.50 | $11.25 |
| Cached input / 1M tokens | $0.06 | $0.13 |
| AA Intelligence Index | 22.8 | 23 |
| AA Coding Index | 52.6 | 37.8 |
| AA Agentic Index | 15.3 | — |
| Max output tokens | 131,072 | 128,000 |
| API providers | 5 | 2 |
| Tool calling | Yes | Yes |
| Reasoning mode | Yes | Yes |
API cost
Estimated cost of 1M input + 1M output tokens via OpenRouter
MiniMax-M2.7
$1.50
GPT-5
$11.25
MiniMax-M2.7 is 7.5× cheaper at list price.
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
MiniMax-M2.7 · fp16 @ 200K ctx
1,111.3 GB / 24 GB