GLM 5.1

Name: GLM 5.1
Author: Zai

Open weights

GLM-5.1 is an AI model from Zai, released with open weights.

Publisher: Zai
Family: Glm
Context: 205K
Input
Output
License: mit
Released: 7 Apr 2026

Cite

Notes

Only stored in your browser.

Attribution

Benchmark scores: frontiermathAA LiveBench prime-hubmatharenaVals AI

Attribution policy →

Context

205K

Input

$1.38

/1M tok

Output

$4.40

/1M tok

Blended

$2.13

#365/ 441/1M · 3:1

Speed

#112/ 441tok/s

Cache read

$0.18

0.13x input

Price & speed via Artificial Analysis · ranked across tracked models · caching via OpenRouter

Reasoning effort	Index	Speed
Reasoning	41.0	0 t/s
Non-reasoning	36.3	0 t/s

Intelligence Index per reasoning-effort setting, via Artificial Analysis · pricing identical across tiers

Reported on 31 evals across 11 domains - best 97.1% on τ²-bench (Tau²-bench) · #9 of 321

Reported eval scores

Graphwalks15.1

AutomationBench35.9

Simpleqa Verified34.6

Skill Reward Hacking5.2

Autonomous Skill Evolution1.1

Apex Shortlist20.8

τ²-bench (Tau²-bench)97.1

MMLU-Pro85.4

Humanity's Last Exam (HLE)27.9

LiveBench72.7

Vals Index52.5

Terminal-Bench (Hard)35.6

LiveBench - Coding75.4

LiveBench - Language71.8

IFBench52

LiveBench - Math84.9

LiveBench - Instruction Following68.5

LiveBench - Data Analysis63.2

LiveBench - Reasoning72.5

GPQA Diamond83.9