TokenRhythm releases NeoHorse-1: 4B/9B models post-trained on agent execution traces
rohanpaul_ai · x · 2026-09-17
rohanpaulai highlights a systemic waste: most agent systems throw away their most valuable training data after every run — which model was routed where, which tools were called, what came back, where the agent failed, and how it recovered.
TokenRhythm's NeoHorse-1 (4B/9B family) is post-trained on agent execution traces and outcomes, i.e. on how tasks were actually executed end to end. Technical report, Hugging Face, ModelScope, and GitHub weights are all available as an open release.
Related event: TokenRhythm Releases NeoHorse-1: Open Models Trained on Agent Trajectories(11 posts)→
More from coding & agent
- Claude Design, Slides and Docs now work inside Claude Code, wired to your repo — ClaudeDevs · 2026-09-17
- Claude Code artifacts evolve into team collaboration: spec docs to implementation — trq212 · 2026-09-17
- Meta's Muse Code coding agent ships natively on Windows, no WSL required — AIatMeta · 2026-09-17
- AI worker startup Delos raises €10M, already live in 300+ companies — iamfakhrealam · 2026-09-17
- A writer built a Claude plugin that learns her voice and improves over time — every · 2026-09-17
- Intent launches free desktop app for spec-driven multi-agent orchestration — LukeW · 2026-09-17