Kimi K2.6vsGLM 5.1
在 9 个共同 benchmark 中，Kimi K2.6 整体领先：Kimi K2.6 领先 8 项，GLM 5.1 领先 1 项，持平 0 项，平均分差 +2.31。
Kimi K2.68 项(89%)(11%)1 项GLM 5.1
评测分数
按能力类目分组，每组内按分差大小排列；共 9 项。
AI Agent - Tool Usage
Kimi K2.6 领先 2/3| 评测项 | Kimi K2.6 | GLM 5.1 | 分差 |
|---|---|---|---|
| Tool Decathlon | 501 / 7Thinking (With Tools) | 40.703 / 7Thinking (With Tools) | +9.30 |
| TerminalBench 2.1 | 53.5613 / 13Thinking (No Tools) | 58.7011 / 13Thinking High (With Tools) | -5.14 |
| Terminal Bench 2.0 | 66.7010 / 46Thinking (With Tools) | 63.5013 / 46Thinking (With Tools) | +3.20 |
General Knowledge
Kimi K2.6 领先 2/2| 评测项 | Kimi K2.6 | GLM 5.1 | 分差 |
|---|---|---|---|
| GPQA Diamond | 90.5016 / 178Thinking (No Tools) | 86.2042 / 178Thinking (No Tools) | +4.30 |
| HLE | 549 / 157Thinking (With Tools + Internet) | 52.3012 / 157Thinking (With Tools) | +1.70 |
Math and Reasoning
Kimi K2.6 领先 2/2| 评测项 | Kimi K2.6 | GLM 5.1 | 分差 |
|---|---|---|---|
| IMO-AnswerBench | 866 / 19Thinking (No Tools) | 83.8010 / 19Thinking (No Tools) | +2.20 |
| AIME 2026 | 96.401 / 14Thinking (No Tools) | 95.302 / 14Thinking (No Tools) | +1.10 |
AI Agent - Information Search
Kimi K2.6 领先 1/1| 评测项 | Kimi K2.6 | GLM 5.1 | 分差 |
|---|---|---|---|
| BrowseComp | 83.2010 / 45Thinking (With Tools + Internet) | 79.3013 / 45Thinking (With Tools + Internet) | +3.90 |
Coding and Software Engineer
Kimi K2.6 领先 1/1| 评测项 | Kimi K2.6 | GLM 5.1 | 分差 |
|---|---|---|---|
| SWE-Bench Pro - Public | 58.607 / 43Thinking (With Tools) | 58.409 / 43Thinking (With Tools) | +0.20 |
规格对比
| 字段 | Kimi K2.6 | GLM 5.1 |
|---|---|---|
| 发布机构 | Moonshot AI | 智谱AI |
| 发布时间 | 2026-04-20 | 2026-03-27 |
| 模型类型 | 推理大模型 | 推理大模型 |
| 架构 | MoE 架构 | MoE 架构 |
| 参数规模 | 1万亿 | 754亿 |
| 上下文长度 | 256K | 200K |
| 最大输出 | 暂无数据 | 125K |
API 调用价格
价格优先使用 DataLearner 配置的 API 记录；缺失项不做推测。
| 价格项 | Kimi K2.6 | GLM 5.1 |
|---|---|---|
| 文本输入 | $0.95 / 1M tokens | $1.4 / 1M tokens |
| 文本输出 | $4 / 1M tokens | $4.4 / 1M tokens |
| 缓存读取 | $0.16 / 1M tokens | $4.4 / 1M tokens |
| 缓存写入 | $0.95 / 1M tokens | $0.26 / 1M tokens |
小结
- Kimi K2.6在以下类目领先:AI Agent - Tool Usage (2/3)、General Knowledge (2/2)、Math and Reasoning (2/2)、AI Agent - Information Search (1/1)、Coding and Software Engineer (1/1)
9 个共同 benchmark 上，Kimi K2.6 平均高出 2.31 分。
单项差距最大的 benchmark：Tool Decathlon — Kimi K2.6 50，GLM 5.1 40.70（分差 +9.30）。
本页正文由结构化模型、价格与 benchmark 数据生成，不使用实时 LLM 撰写。