MiniMax M3 Benchmark Details
MiniMax M3 currently shows benchmark results led by AA-LCR (1 / 27, score 80.33), BrowseComp (12 / 54, score 83.50), SWE-Bench Pro - Public (15 / 59, score 59). This page also compares it with 5 competitor models and 4 predecessor or same-series models, including performance and pricing views when available. 1 source link is attached for reference.
Benchmark Results
Benchmark Results
Coding and Software Engineer
4 evaluationsGeneral Knowledge
2 evaluationsAI Agent - Tool Usage
3 evaluationsProductivity Knowledge
3 evaluationsCompetitor Comparison
Benchmark scores for MiniMax M3 compared against top models in its class
11 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | MiniMax M3Current | GLM-5.2 | Kimi K2.6 | Qwen3.7 Max | GLM 5.1 |
|---|---|---|---|---|---|
GPQA Diamond 科学与综合推理 | 81.31Standard Mode | 91.86Thinking Level · High | 90.50Thinking Enabled | 92.40Thinking Level · High | 86.20Thinking Enabled |
PostTrain Bench 编程与软件工程 | 37.00Thinking Enabled ｜ Tools | 34.30Thinking Level · High ｜ Tools | -- | -- | -- |
SciCode 编程与软件工程 | 45.37Thinking Enabled | -- | -- | 53.50Thinking Level · High | -- |
SWE-Bench Pro - Public 编程与软件工程 | 59.00Thinking Enabled ｜ Tools | 62.10Thinking Enabled ｜ Tools | 58.60Thinking Enabled ｜ Tools | 60.60Thinking Enabled ｜ Tools | 58.40Thinking Enabled ｜ Tools |
Text Arena (Coding) 编程与软件工程 | 1527.75Standard Mode | 1593.25Thinking Level · High | -- | 1540.77Standard Mode | 1534.00Standard Mode |
BrowseComp AI Agent - 信息收集 | 83.50Thinking Enabled ｜ Tools | -- | 83.20Thinking Enabled ｜ Tools | -- | 79.30Thinking Enabled ｜ Tools |
LiveBench 综合评估 | 70.02Deep Thinking Mode | 76.24Standard Mode | 72.17Thinking Enabled | 74.29Deep Thinking Mode | 70.18Standard Mode |
Context Arena 文本向量检索 | 51.15Thinking Enabled | 72.34Thinking Level · High | 64.63Thinking Enabled | 56.01Standard Mode | 62.05Thinking Enabled |
MCP-Atlas AI Agent - 工具使用 | 74.20Thinking Enabled ｜ Tools | 76.80Thinking Enabled ｜ Tools | 69.40Thinking Enabled ｜ Tools | 76.40Thinking Enabled ｜ Tools | 75.60Standard Mode ｜ Tools |
OSWorld-Verified AI Agent - 工具使用 | 70.00Thinking Enabled ｜ Tools | -- | 73.10Thinking Enabled ｜ Tools | -- | -- |
Terminal-Bench 2.1 AI Agent - 工具使用 | 66.00Thinking Enabled ｜ Tools | 81.00Thinking Level · High ｜ Tools | 53.56Thinking Enabled | -- | 58.70Thinking Level · High ｜ Tools |
Standard API Pricing: MiniMax M3 vs. Peer Models
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier.
These models use different currencies or billing units, so the page falls back to raw price values instead of a shared bar chart.
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
MiniMax M3 | MiniMaxAI | ¥2.1 / 1M tokens | ¥8.4 / 1M tokens | — |
GLM-5.2 | 智谱AI | $1.4 / 1M tokens | $4.4 / 1M tokens | — |
Kimi K2.6 | Facebook AI研究实验室 | $0.95 / 1M tokens | $4 / 1M tokens | — |
Qwen3.7 Max | 阿里巴巴 | ¥12 / 1M tokens | ¥36 / 1M tokens | — |
GLM 5.1 | 智谱AI | $1.4 / 1M tokens | $4.4 / 1M tokens | — |
Qwen3.7-Max-Preview | 阿里巴巴 | $2.5 / 1M tokens | $7.5 / 1M tokens | — |
Version History
How each version of the MiniMax M3 series stacks up on benchmark tests
7 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.· Click a row to view its trend chart.
| Benchmark | MiniMax M3Current | MiniMax-M2.7 | MiniMax M2.5 | M2.1 |
|---|---|---|---|---|
GPQA Diamond 科学与综合推理 | 81.31Standard Mode | 87.00Thinking Enabled | 85.20Thinking Enabled | 81.00Thinking Enabled |
SWE-Bench Pro - Public 编程与软件工程 | 59.00Thinking Enabled ｜ Tools | 56.20Thinking Enabled ｜ Tools | 55.40Thinking Enabled ｜ Tools | 32.60Thinking Enabled ｜ Tools |
BrowseComp AI Agent - 信息收集 | 83.50Thinking Enabled ｜ Tools | -- | 76.30Thinking Enabled ｜ Tools | 47.40Thinking Enabled ｜ Tools |
LiveBench 综合评估 | 70.02Deep Thinking Mode | 63.49Deep Thinking Mode | 60.14Deep Thinking Mode | -- |
Context Arena 文本向量检索 | 51.15Thinking Enabled | 33.29Thinking Enabled | -- | -- |
AA-LCR 长上下文能力 | 80.33Thinking Enabled | 69.00Thinking Enabled ｜ Tools | 69.50Thinking Enabled | -- |
GDPval-AA v2 生产力知识 | 1379.75Thinking Enabled ｜ Tools | 1495.00Thinking Enabled ｜ Tools | -- | -- |
Single-Benchmark Version Trend
Viewing: GPQA Diamond · 科学与综合推理
Standard API Pricing Across the MiniMax M3 Series
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier.
These models use different currencies or billing units, so the page falls back to raw price values instead of a shared bar chart.
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
MiniMax M3 | MiniMaxAI | ¥2.1 / 1M tokens | ¥8.4 / 1M tokens | — |
MiniMax-M2.7 | MiniMaxAI | $0.3 / 1M tokens | $1.2 / 1M tokens | — |
MiniMax M2.5 | MiniMaxAI | $0.3 / 1M tokens | $2.4 / 1M tokens | — |
M2.1 | MiniMaxAI | ¥2.1 / 1M tokens | ¥8.4 / 1M tokens | — |