Model dossier模型档案
MiniMax M2.7
MiniMax M2.7 is a high-tier model from MiniMax (Intelligence Score 74.4/100). Strongest in knowledge (86.3) and math (78.3); best fit: knowledge QA and long-form writing. It trails in preference (29.2). Budget-priced: $0.300 in / $1.20 out per 1M tokens. Open weights — self-hostable. Reachable via CN-direct endpoints.
MiniMax M2.7 是 MiniMax 的高水平梯队模型(智能评分 74.4/100)。 最强项是知识(86.3分),其次是数学(78.3分);适合知识问答与长文写作。 短板是偏好(29.2分)。 定价属低价档:每百万 token 输入 $0.300 / 输出 $1.20。 开放权重,可自托管。 支持国内直连。
open weights开放权重 ⌂ CN direct国内直连
74.4
Intelligence Score · benchmark coverage 智能评分 · benchmark 覆盖 93%
◈
Benchmark scoresBenchmark 成绩
10 entries 条| Benchmark基准 | Score成绩 | Index指数 | Dated日期 | Measured by测评方 | |
|---|---|---|---|---|---|
| Humanity's Last ExamReasoning | 29.6 | 52 | — | 3rd-party第三方 | source ↗×4⚠±9.9 |
| GPQA DiamondReasoning | 86.6 | 91 | 2026-03-01 | 3rd-party第三方 | source ↗×6⚠±5.4 |
| MMLU-ProKnowledge | 82 | 86 | — | 3rd-party第三方 | source ↗ |
| SWE-bench VerifiedCoding | 78 | 81 | — | 3rd-party第三方 | source ↗×2 |
| AIME 2025Math | 78.3 | 78 | — | 3rd-party第三方 | source ↗×2 |
| LiveCodeBenchCoding | 79.93 | 85 | — | 3rd-party第三方 | source ↗×2 |
| Terminal-BenchCoding | 57 | 62 | — | official官方榜 | source ↗×3⚠±17.6 |
| τ²-benchAgent | 84.8 | 86 | — | 3rd-party第三方 | source ↗×2 |
| LMArena (Chatbot Arena)Preference | 1416 | 29 | 2026-08-12 | official官方榜 | source ↗×2 |
| WebArenaAgent | 63.1 | 68 | — | 3rd-party第三方 | source ↗ |
≡
Specs & pricing规格与价格
| Provider厂商 | MiniMax |
|---|---|
| Released发布 | 2026-03-18 |
| Context window上下文 | 205K tokens |
| Max output最大输出 | 131K tokens |
| Modality模态 | text->text |
| Reasoning tiers推理档位 | not documented未见公开文档 |
| API input price输入价格 | $0.300 / 1M tokens |
| API output price输出价格 | $1.20 / 1M tokens |
How we score →评分方法 → · Catalog & pricing via the official OpenRouter API.模型目录与价格来自 OpenRouter 官方 API。