Google ships a decision-making AI model into Chrome, tested against Gemini Nano and Decisions API
gaganghotra_ · x · 2026-10-09
Sharing @natzir9's 9-part benchmark thread: Google is embedding AI models into Chrome, the latest being a model purpose-built for decisions. The author tested it on their own site against Gemini Nano, OpenAI's Decisions API, and Jev.
Early takeaway: the smallest model isn't the most accurate, but it knows when it doesn't know — it can abstain, which may matter more than raw accuracy for decision tasks. Full numbers in the thread.
More from Models
- Dev racing to burn $800 of Gemini API credit teases rumored Gemini 4 Argon and Nano Banana Pro 2 — Angaisb_ · 2026-10-09
- 16 VPD weight edits boost model accuracy 10x, revealing the attention head that suppresses introspection — Sauers_ · 2026-10-09
- GPT-6.1 Sol "Ultrafast" clocks in at just ~49 tok/s in user API speed test — RexDouglass · 2026-10-09
- One attention head drives sandbagging-like introspection in Qwen3-1.7B; ablating it helps — Sauers_ · 2026-10-09
- Text-Only Qwen3.5 2B/4B/9B MLX 4-bit Packages Released, 2B Is Just 1GB — sachasayan · 2026-10-09
- Why multilingual LLMs are hard: character decoding and BPE are the hidden bottleneck — ivan_bezdomny · 2026-10-09