evals.report
BenchmarksLabsCompareRun guidesIn the wild
Z.aiGLM-4 (GLM-4.7 series)

GLM-4.7

Z.ai · GLM-4 (GLM-4.7 series). Released Dec 22, 2025.

GLM-4.7 is a model from Z.ai in the GLM-4 (GLM-4.7 series) family, released Dec 22, 2025. evals.report tracks 16 reported GLM-4.7 benchmark scores across SWE-bench Verified, GPQA Diamond, Humanity's Last Exam, PostTrainBench, Artificial Analysis Intelligence Index, Epoch Capabilities Index, SWE-rebench, MMLU-Pro, and 8 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

Open16 results

Benchmark results 16

Compare this model
BenchmarkCategoryScoreMetricStatusDate
SWE-bench VerifiedCoding73.8%% resolvedVerifiedDec 22, 2025Details
GPQA DiamondReasoning85.7%accuracyVerifiedDec 22, 2025Details
Humanity's Last ExamReasoning24.8%accuracyVerifiedDec 22, 2025Details
PostTrainBenchAgents7.48%weighted average scoreOfficialDec 22, 2025Details
Artificial Analysis Intelligence IndexReasoning42.1IndexUnverifiedDec 22, 2025Details
Epoch Capabilities IndexReasoning144.6IndexOfficialDec 22, 2025Details
SWE-rebenchCoding58.7%Resolved rate (pass@1)UnverifiedDec 22, 2025Details
MMLU-ProReasoning85.6%accuracyVerifiedDec 22, 2025Details
GDPvalAgents1185EloOfficialDec 22, 2025Details
SciCodeCoding45.1%accuracyUnverifiedDec 22, 2025Details
Global-MMLUReasoning79.9%accuracyUnverifiedDec 22, 2025Details
WebDev ArenaChat preference1440EloVerifiedDec 22, 2025Details
EQ-Bench Creative Writing v3Chat preference1403EloVerifiedDec 22, 2025Details
Design ArenaChat preference1273EloVerifiedDec 22, 2025Details
Vectara Hallucination LeaderboardOther11.7%Hallucination RateOfficialDec 22, 2025Details
Terminal-Bench 2.0Agents41.0%task successVerifiedDec 22, 2025Details