NeoHorse-1 open-sourced: 4B/9B Qwen3.5 post-trained models targeting recursive self-improvement
PrajwalTomar_ · x · 2026-09-16
TokenRhythm released NeoHorse-1, an open-source project pitched as an early prototype toward recursive self-improvement (RSI), with a technical report on arXiv.
Key points:
- Ships 4B and 9B checkpoints post-trained from Qwen3.5 for agent harnesses, tool use, coding, and instruction following.
- The core is a routing harness that assigns tasks to a heterogeneous model pool, logs tool interactions and outcomes, estimates capability demand, and feeds capability-level feedback into the next training mixture.
- Updated models re-enter the harness, forming an evaluation–selection–update loop; extending it across iterations is the next step toward RSI.
- The repo has 400+ stars and offers GGUF quantizations. The author argues the next AI edge comes from owning the learning loop, not chasing one perfect model.
More from Models
- Indie dev open-sources Qwen-2.5-1B-RLCD with 5x faster on-device JSON inference — suchenzang · 2026-09-16
- LLaDA-Image: 6B fully-diffusion DiT trained on 90% image-only data, no caption bottleneck — jiqizhixin · 2026-09-16
- GPT-5.6 Sol flips its conclusions when you just ask 'Are you sure?' — Sockand2 · 2026-09-16
- Student finds Opus and ChatGPT trash his slides, then reverse after seeing the source material — Kazoru4 · 2026-09-16
- Altman says internal post-Astra OpenAI model can solve problems the world's best mathematicians cannot — rohanpaul_ai · 2026-09-16
- Agent randomly references a money quote the user never said while editing unrelated text — Kyrannio · 2026-09-16