GLM-5.3 · Agents' Last Exam (ALE-CLI)(GLM-5.3模型卡;ALE-CLI) 28.5 · 主文证据(ev-glm53-ale-cli)

GLM-5.3 的 Agents' Last Exam (ALE-CLI) · GLM-5.3模型卡;ALE-CLI 官方出处。GLM-5.3 · max · 28.5。见《GLM-5.3 model card | zai-org》Benchmark table, Agents' Last Exam (ALE-CLI) row, GLM-5.3 column; Agents' Last Exam footnote。证据 ev-glm53-ale-cli。

原文与定位

GLM-5.3 · Agents' Last Exam (ALE-CLI)(GLM-5.3模型卡;ALE-CLI),GLM-5.3 · max配置,Reported score 28.5(具体成绩公布时间未确认)。见《GLM-5.3 model card | zai-org》Benchmark table, Agents' Last Exam (ALE-CLI) row, GLM-5.3 column; Agents' Last Exam footnote。 查看成绩条件

| Agents' Last Exam (ALE-CLI) | 28.5 | 23.8 | 27.6 | 25.7 | 27.0 | 25.7 | 23.8 | **28.6** | **Agent's Last Exam (CLI)**: We evaluate ALE using the official evaluation protocol with the Claude Code harness (reasoning effort=max, 1M context, and 64K maximum output). Each of the 105 tasks runs in an isolated Docker container using the resources declared in its Task Card. The default timeout is 4 hours, with task-specific limits taking precedence (up to 8 hours). Tool Search is disabled, and results are scored by the official ALE evaluators.
官方材料
GLM-5.3 model card | zai-org
定位
主文 · Benchmark table, Agents' Last Exam (ALE-CLI) row, GLM-5.3 column; Agents' Last Exam footnote
材料版本
官方 Hugging Face main 分支模型卡;表格、评测脚注与配置说明已核读;2026-11-17 采集
版本公布
未确认
披露时间
未确认 · 首发/后补未确认
采集时间
历史依据
暂无可靠历史版本依据
引用文本

GLM-5.3 · Agents' Last Exam (ALE-CLI)(GLM-5.3模型卡;ALE-CLI),GLM-5.3 · max配置,Reported score 28.5(具体成绩公布时间未确认)。官方出处 GLM-5.3 model card | zai-org(主文,官方 Hugging Face main 分支模型卡;表格、评测脚注与配置说明已核读;2026-11-17 采集)Benchmark table, Agents' Last Exam (ALE-CLI) row, GLM-5.3 column; Agents' Last Exam footnote,https://huggingface.co/zai-org/GLM-5.3/blob/main/README.md。Benchlog https://benchlog.lekshome.cc/evidence/glm-5.3/ev-glm53-ale-cli/

关联成绩与条件

GLM-5.3 · Agents' Last Exam (ALE-CLI) 成绩与配置
指标与成绩配置与评测条件证据关系
Reported score28.5原文 28.5 · 单位 未标注
公布 未确认
GLM-5.3 · max
设置
官方模型卡目标列;逐行scaffold、工具、采样及预算条件见protocol。来源把benchmark复现明确绑定max或行内明确reasoning_effort=max。
未知项
未逐项说明的指标单位、题集修订、服务精度和运行次数保持未知。
工具
工具条件未记录
推理预算
max · 输出上限 64K tokens
框架
Claude Code
协议
The card says reasoning_effort defaults to max and recommends keeping that default for benchmark reproduction. **Agent's Last Exam (CLI)**: We evaluate ALE using the official evaluation protocol with the Claude Code harness (reasoning effort=max, 1M context, and 64K maximum output). Each of the 105 tasks runs in an isolated Docker container using the resources declared in its Task Card. The default timeout is 4 hours, with task-specific limits taking precedence (up to 8 hours). Tool Search is disabled, and results are scored by the official ALE evaluators.
条件
GLM-5.3; task release unspecified. The GLM source does not state a unit; no percentage inferred. 2026-11-18按完整模型卡修正单位、指标/题集或协议标注,数字不变。

直接支持该成绩

在排名中查看此成绩