AnalysisAI ModelsNovember 3, 2026

Axion 1.0: open agent foundation model with a three-stage recipe

Read original source →arxiv.org

Axion 1.0 is a fully open agent foundation model trained in three stages — two lightweight supervised fine-tuning passes then reinforcement learning — with data and training recipe released. The 35B weights are on Hugging Face; the paper reports competitive agentic performance on tool-dependent and long-horizon reasoning tasks.

2 sources

More stories today

Open the live feed