Benchmark Almanac

Z.ai

Best on the Capabilities Index GLM-5.3 156.2
27 models 16 model cards read z.ai

Z.ai's models against the field

Every model on the Epoch Capabilities Index by release date, with Z.ai's in color.

50 75 100 125 150 175 2023 2024 2025 2026 2027 GLM-5.3-Flash GLM-5 50 75 100 125 150 175 2023 2024 2025 2026 2027 GLM-5.3-Flash GLM-4.7

Its strongest models

Z.ai Aug 2026

GLM-5.3

Open weights Flagship

156.2 #33 of 271
Epoch Capabilities Index
  • DeepSWE 69% #7 /28
  • NL2Repo-Bench 58% #6 /28
  • Agents' Last Exam 26.2% #11 /12
Z.ai Aug 2026

GLM-5.3-Flash

Open weights

153.2 #48 of 271
Epoch Capabilities Index
  • DeepSWE 63.4% #13 /28
  • NL2Repo-Bench 56.3% #8 /28
  • Terminal-Bench 4.0 35.8% #12 /23
Z.ai Jun 2026

GLM-5.2

Open weights Flagship

152.1 #53 of 271
Epoch Capabilities Index
  • SWE-bench Verified 78.7% #5 /32
  • DeepSWE 43.8% #24 /28
  • NL2Repo-Bench 48.9% #11 /28
Z.ai Apr 2026

GLM-5.1

Open weights Flagship

149.9 #65 of 271
Epoch Capabilities Index
  • SWE-bench Verified 74.2% #16 /32
  • Toolathlon 40.7% #18 /50
  • NL2Repo-Bench 42.7% #18 /28
Z.ai Feb 2026

GLM-5

Open weights Flagship

145.8 #89 of 271
Epoch Capabilities Index
  • SWE-bench Verified 72.1% #22 /32
  • IFBench 72% #33 /70
  • Terminal-Bench 2.0 52.4% #16 /43
Z.ai Dec 2025

GLM-4.7

Open weights Flagship

143.7 #102 of 271
Epoch Capabilities Index
  • Terminal-Bench 2.0 33.4% #30 /43
  • Toolathlon 23.8% #37 /50
  • CritPt 1.71% #86 /160

What its model cards measure

Tall ticks are Z.ai model cards that list the benchmark in their headline table. Gray dots are cards that leave it out. Rows are the benchmarks the lab listed most, at least twice each, in the order it started using them.

GPQA Diamond AIME 2025 OCRBench MMBench-EN v1.1 MMBench-CN v1.1 MMStar BLINK MuirBench MMMU MMMU-Pro Video-MMMU AI2D MathVista We-Math ChartQA Pro ChartMuseum MMLongBench-Doc RefCOCO / RefCOCO+ / RefCOCOg OSWorld AndroidWorld Humanity's Last Exam BrowseComp SWE-bench Verified tau2-bench HMMT Nov 2025 IMO-AnswerBench Toolathlon NL2Repo-Bench

Model cards

DateModelDocumentBenchmarks
26 Aug 2026 GLM-5.3-Flash GLM-5.3-Flash: Frontier Intelligence, Flash Cost (blog) 16
14 Aug 2026 GLM-5.3 GLM-5.3 model card (Hugging Face) 16
16 Jun 2026 GLM-5.2 GLM-5.2 model card (Hugging Face) 17
7 Apr 2026 GLM-5.1 GLM-5.1 model card (Hugging Face) 15
11 Feb 2026 GLM-5 GLM-5 model card (Hugging Face) 15
30 Jan 2026 GLM-OCR GLM-OCR model card (Hugging Face) 2
19 Jan 2026 glm-4.7-flash GLM-4.7-Flash model card (Hugging Face) 7
22 Dec 2025 GLM-4.7 GLM-4.7 model card (Hugging Face) 15
8 Dec 2025 GLM-4.6V GLM-4.6V blog and model card 32
30 Sep 2025 GLM-4.6 GLM-4.6 launch blog (Z.ai) and Hugging Face card, chart 'LLM Performance Evaluation: Agentic, Reasoning and Coding' 8
11 Aug 2025 GLM-4.5V GLM-4.5V model card (Hugging Face) 40
28 Jul 2025 GLM-4.5 GLM-4.5: Reasoning, Coding, and Agentic Abilities (Z.ai launch blog) 13
1 Jul 2025 GLM-4.1V-9B-Thinking GLM-4.1V-9B-Thinking model card and technical report 26
14 Apr 2025 GLM-Z1-32B-0414 GLM-4-0414 model series README (GitHub) 10
16 Jan 2024 GLM-4 GLM-4 launch at Zhipu DevDay (official announcement text and slides, reposted on Volcengine developer community) 10
27 Oct 2023 ChatGLM3-6B ChatGLM3 GitHub README, base-model evaluation table 8

Every Z.ai model