evals.report
BenchmarksLabsCompareRun guidesIn the wild
Z.aiGLM

GLM-5.1

Z.ai · GLM. Released Apr 7, 2026.

GLM-5.1 is a model from Z.ai in the GLM family, released Apr 7, 2026. evals.report tracks 27 reported GLM-5.1 benchmark scores across FrontierMath, DeepSWE, LMArena, LiveBench, GPQA Diamond, SWE-bench Verified, Humanity's Last Exam, AIME (OTIS Mock), and 19 more — each shown with its benchmark, metric, source status, and date, and never combined into a single ranking.

Open27 results

Benchmark results 27

Compare this model
BenchmarkCategoryScoreMetricStatusDate
FrontierMathReasoning33.45%accuracyOfficialApr 7, 2026Details
DeepSWECoding17.48%% resolvedOfficialApr 7, 2026Details
LMArenaChat preference1469source-defined ratingOfficialApr 7, 2026Details
LiveBenchReasoning70.18%scoreOfficialApr 7, 2026Details
GPQA DiamondReasoning85.5%accuracyOfficialApr 7, 2026Details
SWE-bench VerifiedCoding74.2%% resolvedOfficialApr 7, 2026Details
Humanity's Last ExamReasoning25.63%accuracyOfficialApr 7, 2026Details
AIME (OTIS Mock)Reasoning92.2%accuracyOfficialApr 7, 2026Details
SimpleQA VerifiedOther37.3%accuracyOfficialApr 7, 2026Details
WeirdMLCoding57.1%average accuracyOfficialApr 7, 2026Details
Artificial Analysis Intelligence IndexReasoning51.4IndexUnverifiedApr 7, 2026Details
Epoch Capabilities IndexReasoning149.9IndexOfficialApr 7, 2026Details
SWE-rebenchCoding62.7%Resolved rate (pass@1)OfficialApr 7, 2026Details
τ²-bench (Telecom)Tool use97.7%pass^1OfficialApr 7, 2026Details
AIME 2026Reasoning95.83%accuracyOfficialApr 7, 2026Details
GDPvalAgents1535EloOfficialApr 7, 2026Details
SciCodeCoding43.8%accuracyUnverifiedApr 7, 2026Details
AA-Omniscience: Knowledge and Hallucination BenchmarkReasoning2AA-Omniscience IndexOfficialApr 7, 2026Details
IFBenchReasoning76.3%accuracyOfficialApr 7, 2026Details
WebDev ArenaChat preference1533EloVerifiedApr 7, 2026Details
EQ-Bench Creative Writing v3Chat preference1645EloVerifiedApr 7, 2026Details
Design ArenaChat preference1335EloVerifiedApr 7, 2026Details
MathArena HMMT February 2026Reasoning89.39%accuracyOfficialApr 7, 2026Details
FrontierMath Tier 4Reasoning12.5%accuracyOfficialApr 7, 2026Details
Vibe Code BenchCoding31.46%Overall accuracyVerifiedApr 7, 2026Details
Terminal-Bench 2.0Agents63.5%task successVerifiedApr 7, 2026Details
FrontierSWEAgents31%dominance scoreOfficialApr 7, 2026Details