Stanford and NVIDIA's CLM embeds decisions instead of tokens, 13x faster but accuracy drops at scale
Prompt Engineering · youtube · 2026-09-26
CLM: embedding-match instead of token generation
Stanford and NVIDIA Research introduced Contrastive LM (CLM), a "System 1" model that works like CLIP for decisions: it embeds the situation and every candidate action into one space and picks the nearest match instead of generating tokens, making decisions 13x faster.
Training and benchmarks:
- Built on a frozen 8B backbone with only small heads being trained
- Tested on a DGX Spark across 1,080 tools: latency held around 80ms, but accuracy fell from 86% with 8 tools to 17% with all of them
Positioning: not a replacement for reasoning models today, but a fast first-stage shortlist before a model that compares options side by side.
Code, weights (8B) and a playground are open-sourced.
More from Infra
- Why rent servers when agents can run your terminal? — StewartalsopIII · 2026-09-26
- MIT paper: scaling law expiring as cost doubles 6x per width jump — DavidLinthicum · 2026-09-26
- Investor thesis: GOES steel and transformer makers may outshine rare earths — basedjensen · 2026-09-26
- Polymorf pushes OMLX inference from 150tps to nearly 200tps within 24 hours of launch — HankYeomans · 2026-09-26
- FlashLoop exploits cross-loop redundancy to speed up Looped Transformers by 1.65x — KyeGomezB · 2026-09-26
- Why this builder quit server racks: fried motherboards and a ~$1,500 housing bill — TheZachMueller · 2026-09-26