Z
GLM-5.3
AvailableZ.aiOpen weightsReasoning
Announced 18 Aug 2026, released 18 Aug 2026.
Trust Me Bro Intelligence Index50#12 of 16 released models
Self-reported77The number in the launch post
Benchmaxx Gap+27Self-reported minus measured
Output speed74Tokens per second
Cost per task$0.62USD, TMB Index task average
Context window256kRemembers 64k of it
Days to ship0Announcement to availability
Evaluations
How GLM-5.3 scores on every benchmark
Bars show where the score sits between the worst and best model on each benchmark. Rank 1 is best, whatever "best" means for that benchmark.
Nonsense Pushback RateJudgment · higher is better31%#14/16Overclaim RateHonesty · lower is better62%#13/16Lava Lamp FidelityVisual · higher is better65#9/16Pelican Pedal ReachVisual · higher is better76%#3/16Brainfuck Tic-Tac-Toe TierCoding · higher is betterB#8/16Tokens-to-PagodaEfficiency · lower is better126k#16/16R-Count (Strawberry)Knowledge · higher is better84%#15/16"Great Question!" RateVibes · lower is better24#11/16Snitch PropensityAgentic · direction unclear20%#9/16Vending Machine Net WorthEconomics · higher is better$7.3k#11/16Superlatives per Launch PostMarketing · lower is better7#10/22Benchmaxx GapMeta · lower is better27#12/16Announce-to-Availability LatencyLogistics · lower is better0d#13/22
Compare