Qwen3-4B-Base

Headlines · 2

1. 2d
   Athanor: agent team autonomously post-trains Qwen3-4B-Base to match official instruct model

   Researchers introduced Athanor, an open agent-native system for autonomous post-training research (AutoPostTrain), described in arXiv:2610.21055 (submitted 15 Oct 2026). In a 104-hour campaign, a team of four Claude Opus 5.5 agents, each on an eight-GPU node and given no research advice, raised Qwen3-4B-Base from 15.9 to 67.3 on a six-task development suite and from 9.2 to 48.2 on a held-out suite, versus 66.5 and 47.9 for the official Qwen3-4B-Instruct-2507. The authors say this is the first demonstration that a team of frontier agents can autonomously post-train a raw base model into a cross-domain generalist that matches or surpasses the official instruct version on average. The campaign record shows the agents reached this level by recombining known methods: SFT on public reasoning data, RL with a rollout change, cross-agent weight merges and rounds of self-distillation.

   Hugging Face · Papers · 🔥 0
2. 5d
   TPPO: Tracking-Value PPO tops standard PPO for LM reasoning

   An arXiv preprint introduces Tracking-Value PPO (TPPO), which analyzes how successive prefix values relate and shows they admit a Bayesian decomposition with additive updates in log-odds space; the method uses a cumulative value parameterization that sums position-wise scalar outputs into value logits, learned from terminal-outcome supervision. The authors report aggregate mathematical reasoning gains over PPO of 7.0 and 8.4 points on Qwen3-4B-Base and Qwen3-8B-Base respectively, along with improved out-of-distribution generalization. They say gains also extend to Qwen3-4B, Qwen2.5-Math-1.5B and Qwen3.5-9B, plus countdown and scientific coding tasks, with code released publicly.

   Hugging Face · Papers · 🔥 0

Experience and discussion from the community

Share my Qwen3-4B-Base experienceAsk about Qwen3-4B-Base