Frontier releases suggest open models like Kimi K3 are severely undertrained
zeeshanp_ · x · 2026-09-23
Drawing on recent frontier model releases, the author argues that many large open models such as Kimi K3 are likely severely undertrained, implying that data scaling still has a long way to go and the open-vs-frontier gap partly stems from insufficient training data.
More from Infra
- Fully Offline NotebookLM Alternative: Ollama + Open WebUI + RAG Stack Suggested — betobagio · 2026-09-23
- Critic calls Googlebook hardware 'unambitious': Intel opted out of consumer NPUs — julianharris · 2026-09-23
- vLLM ships nightly support for Google's DiffusionGemma, a block-diffusion MoE with ~1.9x throughput — vllm_project · 2026-09-23
- 65-70% of LLM speedup claims unimpressive, says dev: it's mostly speculative decoding and prompt lookup — teortaxesTex · 2026-09-23
- ZeroHedge claims OpenAI Stargate's two key data centers hit funding snags — ns123abc · 2026-09-23
- DarkbloomAI one month paid on OpenRouter: 4B to 20B+ tokens/day — gajesh · 2026-09-23