GLM-5.3 · PostTrainBench(GLM-5.3模型卡;unknown) 39.8 · 主文证据(ev-glm53-posttrainbench)

GLM-5.3 的 PostTrainBench · GLM-5.3模型卡;unknown 官方出处。GLM-5.3 · max · 39.8。见《GLM-5.3 model card | zai-org》Benchmark table, PostTrainBench row, GLM-5.3 column; PostTrainBench footnote。证据 ev-glm53-posttrainbench。

原文与定位

GLM-5.3 · PostTrainBench(GLM-5.3模型卡;unknown),GLM-5.3 · max配置,Reported score 39.8(具体成绩公布时间未确认)。见《GLM-5.3 model card | zai-org》Benchmark table, PostTrainBench row, GLM-5.3 column; PostTrainBench footnote。 查看成绩条件

| PostTrainBench | 39.8 | 31.7 | 32.0 | – | – | 32.9 | **41.8** | 36.2 | **PostTrainBench**: We evaluate GLM-5.3 using Claude Code 2.1.207 with max effort level, `temperature = 1.0`, `top_p = 1.0`, `max_new_tokens = 128000`, and a 1M-token context window. We report the weighted average over 3 runs. Runs that fail to produce a score fall back to the official zero-shot base-model baseline score. For checks intended to prevent the use of third-party APIs, we removed the original pattern-matching-based checks, as they produced false positives when a local vLLM endpoint was accessed through the OpenAI SDK. Instead, we use an LLM agent to inspect solutions for external API usage.
官方材料
GLM-5.3 model card | zai-org
定位
主文 · Benchmark table, PostTrainBench row, GLM-5.3 column; PostTrainBench 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 · PostTrainBench(GLM-5.3模型卡;unknown),GLM-5.3 · max配置,Reported score 39.8(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, PostTrainBench row, GLM-5.3 column; PostTrainBench footnote,https://huggingface.co/zai-org/GLM-5.3/blob/main/README.md。Benchlog https://benchlog.lekshome.cc/evidence/glm-5.3/ev-glm53-posttrainbench/

关联成绩与条件

GLM-5.3 · PostTrainBench 成绩与配置
指标与成绩配置与评测条件证据关系
Reported score39.8原文 39.8 · 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 128000 tokens
框架
Claude Code 2.1.207
协议
The card says reasoning_effort defaults to max and recommends keeping that default for benchmark reproduction. **PostTrainBench**: We evaluate GLM-5.3 using Claude Code 2.1.207 with max effort level, `temperature = 1.0`, `top_p = 1.0`, `max_new_tokens = 128000`, and a 1M-token context window. We report the weighted average over 3 runs. Runs that fail to produce a score fall back to the official zero-shot base-model baseline score. For checks intended to prevent the use of third-party APIs, we removed the original pattern-matching-based checks, as they produced false positives when a local vLLM endpoint was accessed through the OpenAI SDK. Instead, we use an LLM agent to inspect solutions for external API usage.
条件
GLM-5.3; benchmark version, score definition and unit unspecified. 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