Kimi K3 Effective Parameters Criticized: Only Matching 2022's PaLM-1
teortaxesTex · x · 2026-07-28
Tech commentator @teortaxesTex estimated the effective parameter count of Moonshot's newly released Kimi K3 using the sqrt(activetotal) rule of thumb, yielding around 539B. He sharply noted that this merely reaches the level of Google's PaLM-1 from April 2022, suggesting a significant gap still exists for Chinese LLMs.
More from Models
- Kimi K3 lands near the top of Artificial Analysis’s intelligence index — ccerrato147 · 2026-07-28
- Qwen 3.6 27B agent gets much smarter after KV cache quantization change — Jordanthecomeback · 2026-07-28
- NYT Reports on Kimi K3: Chinese AI Startup Shows the World What It Has Built — pstAsiatech · 2026-07-28
- Kimi K3 turns a prompt into a playable Chrome Dino game, then debugs itself — BanghuaZ · 2026-07-28
- A second Opus 5 prompt renders another highly detailed Pip-Boy UI — chrisfirst · 2026-07-28
- Kimi K3 lands day-one on Dell PowerEdge XE9780 for on-prem deployment — _akhaliq · 2026-07-28