Epoch AI: open-weight models trail closed frontier on cyber by 6–7 months
morqon · x · 2026-10-17
Epoch AI estimates open-weight models lag the closed frontier by about 6–7 months on cyber capabilities (Cyber ECI). Zai's GLM-5.3 is the highest-scoring open-weight model at 160, beating OpenAI's GPT-5.4 released in March.
Dean Ball adds: during AI-enabled research acceleration, a consistent 6–9 month lag in objective time means the real gap keeps widening—today's 6-month lag is subjectively larger than one from 6 months ago.
More from Models
- DeepSeek flash 1017 called out for still acting "very much like Claude" — teortaxesTex · 2026-10-17
- Anthropic says distillation defenses cut harvested data 90-99%, names Zhipu GLM-5.3 — bibryam · 2026-10-17
- JevBench Adds 34 Decision Models in 4 Days as kettei-fat Takes #1 — airesearch12 · 2026-10-17
- OpenAI's Ultrafast tier drops to ~65 tok/s, breaking paid priority routing plans — thepriceisright__ · 2026-10-17
- Diffusion Gemma promises 4x faster text generation; may supplant autoregressive LLMs — NandoDF · 2026-10-17
- Perplexity reveals pplx-decider: reranker with zero reranking supervision, private MTEB-style eval — bo_wangbo · 2026-10-17