2026-10-11 18:00 UTC

small-language-models

band: coolmomentum: stable score: 0.003
temperature history

Episodes (4)

Independent evaluation will determine whether Unbounded Labs’ released 2.82B-parameter Bart model, trained from scratch on 20.1B tokens of pre-1931 English, provides a useful testbed for historical-language behavior and novel-idea generation.
expiredknownscott: low
Independent evaluations will determine whether Liquid AI’s LFM2.5-2.6B enables practically useful agent workloads on edge and resource-constrained hardware.
resolvedknownscott: medium
Neurometric claims its task-specific small language model delivers a useful quality-cost tradeoff for tool calling, potentially making narrow agent workflows cheaper to operate.
expiredknownscott: low
Thomson Reuters claims Thomson-1.0-Small is a 35.1B-parameter MoE model specialized for legal, tax, journalism, and tool-using research workflows, potentially giving professional deployments a publisher-trained domain model with long context.
seedknownscott: low

Trajectory notes