GLM 5.3 "Flash" spotted internally; observers doubt it's a small model
teortaxesTex · x · 2026-08-21
A user claims Zhipu's GLM 5.3 "Flash" has been spotted internally. teortaxesTex argues it may not actually be a smaller Flash-tier model: first, it's slow for a Flash model (though with the 100T compute commitment it could be run at high batch size); second, its knowledge seems on par with 5.3 proper.
He speculates that if it's under 740B parameters, Zhipu surpasses DeepSeek (and Kimi) by a margin that's "not even funny." (Unverified internal sighting and speculation.)
More from Models
- Gemini 3.7 Flash tops ARC-AGI benchmark at $0.12 per task — rakyll · 2026-08-21
- o1 excels at fixing vision-grounded bugs in rendering pipelines — teortaxesTex · 2026-08-21
- Mystery stealth model 'Ox Alph' appears on OpenRouter, China lab suspected — Neosinic · 2026-08-21
- Claude Output Occasionally Mixed with Chinese Characters — springrod · 2026-08-21
- GLM-5.3 Vision questioned as Ox Alpha source; model scale deemed non-essential — teortaxesTex · 2026-08-21
- Why people imagine Grok would have full CoT visibility — teortaxesTex · 2026-08-21