GLM-4.7 Flash (Non-reasoning) — benchmark results

GLM-4.7 Flash evaluated with reasoning disabled. Provider: Zhipu. Released 2026-01-19. Access: Open.

Unified ELO 1565 ± 43, rank #579 of 1776 rated models, from 35 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA TAU-2 Bench91.81Accuracy (%)86.9
UGI - Willingness (W/10)6.8W/10 Score66.6
AA IFBench46.26Accuracy (%)54.4
Artificial Analysis Intelligence Index15.52Intelligence Index50.4
AA Omniscience - Software Engineering (SWE) - Julia12Accuracy (%)47.9
UGI Leaderboard33.75UGI Score47.2
UGI - Natural Intelligence20.96NatInt Score45.8
AA Omniscience - Software Engineering (SWE) - Java15Accuracy (%)41.3
AA Omniscience - Software Engineering (SWE) - Kotlin12Accuracy (%)31.1
AA SciCode25.46Accuracy (%)30.4
AA Humanity's Last Exam4.87Accuracy (%)30.1
AA Omniscience - Software Engineering (SWE) - Python14Accuracy (%)28.9

Interactive version: theaggregate.ai/model?slug=glm-4-7-flash-non-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-22.