Models
Why Mistral Large 4's blog post and model card disagree on its size
Mistral's 2026-10-06 announcement says 49B active and 1T total parameters; its model card says 52B active and 1.05T total.
Whistle puts speech-to-text for on-device agents in 16.9 MB
Cactus Whistle packages 7-language speech recognition into a 16.9 MB CPU file that pairs with Needle to emit structured tool calls directly from audio.
OpenAI will watermark Codex and ChatGPT text in the EU
OpenAI will add invisible statistical watermarks to eligible ChatGPT and Codex text in the EU, while API customers worldwide can opt in for select models.
What a System One model decides, and what it refuses
System One models return typed probabilities for bounded questions and refuse free-form text, code, and reasoning explanations.
Clef, Clef-flash, or Jev: which decision model belongs on the agent hot path?
Use Clef-flash for speed and vision, Clef for a larger model and context window, or Jev for low-cost text routing.
Liquid's d1 now answers from an image
Liquid AI added vision to its d1 decision model on 2026-10-05, returning calibrated probabilities at $0.04 per million input tokens with zero output tokens.
Reflection Beam: 501B open-weight model, weights later in October 2026
Reflection Beam is a 501B sparse MoE with 23B active parameters for coding and agent tasks; Reflection plans Apache 2.0 weights later in October 2026.
GPT-5.6 Sol, Terra, and Luna: which tier to use for agent work
Use Sol for multi-agent reasoning and long context, Terra as the default for coding at 40% of the cost, and Luna only for short, high-volume tasks.