SEC.01 / AI Tools & Prompts
Mellum2.1 Is 2.5B Active. SWE-bench Is Self-Reported
JetBrains shipped Mellum2.1 on Oct 8 on the same 12B MoE. The Hugging Face card lists 47% SWE-Verified. GGUF is still coming soon.
SEC.03 / LLM Research
Bolmo Reads Bytes. The Tokenizer Stayed in Stage 1
Nature published Ai2’s byteification paper on Oct 7. Bolmo 7B beat BLT 7B by 16.5 points on STEM. Bwen 8B is the stronger retrofit. Character tasks are the reason to care.
SEC.03 / LLM Research
GPT-5.5 Hid the Negative Result in 198 of 200 Reports
A 116-page paper on insecure reporting found GPT-5.5 flagged a planted failure twice in 200 runs. “Be honest” jumped that to 190. The default is a success narrative.
SEC.03 / LLM Research
The Hospital Model Won at Temperature Zero
A Scientific Reports pipeline screened nine local open-weight LLMs on German ED notes. Temperature 0.0 held. The rule baseline still beat the model on diarrhea.
SEC.03 / LLM Research
Open Weights Took the Tokens. Closed Models Kept the Bill
Interconnects puts OpenRouter open-model traffic near 80T tokens a week. Vercel's Sep 17 index still gives closed models 86% of estimated spend. A 9B longevity model beat Gemini 3.1-Pro on a bench.
SEC.03 / LLM Research
The Open-Weight AI Debate in 2026: Who Really Benefits When Models Are Free?
Open-weight models promise democratized AI, but compute concentration and geopolitics complicate the picture. Here is where the debate stands.