Model · Z.ai

Z.ai Glm 5.2

3 appearances · meta-score 69.9/100 · last seen 2026-07-17 · scored under rubric v15

Pooled composite across 3 primary runs: 54.3 ± 18.3 (95% CI)

$75.20Total cost
127,046,924Tokens
77.0%Cache hit
2,026Requests
0.25%Failure rate
1.19Tools/response
1,689,460Tokens per $

Agent role scores

Per-agent-role scores for Z.ai Glm 5.2
Agent role	Score / 10
Sprint-Review Agent	8.57
Orchestrator	7.64
Scout Agent	7.60
Develop Agent	7.56
Build Agent	7.19
Sprint-Plan Agent	6.96
Code-Review Agent	6.93
Code Agent	6.42
Test Agent	6.25
Pseudocode Agent	5.39

Run history (primary model)

Runs where Z.ai Glm 5.2 was the primary model, oldest first
Run	SoW	Date	Composite
Glm 5.2	circles	Jun 2026	46.9
Glm 5.2	circles	Jul 2026	54.5
Glm 5.2	circles	Jul 2026	61.6

Every run on Favur Evals is scored by the same deterministic engine, on the same Statements of Work. The benchmark is self-funded — no vendor sponsorship, credits, or grants.