Kardashev-0.7: Banbury trains 32 distinct models together with reinforcement learning
Summary
Banbury Road unveils Kardashev-0.7, a population of 32 distinct models trained jointly with reinforcement learning to pursue Population Scaling, treating model count and collaboration as a new scaling axis toward complementary specializations.
Similar Articles
@alex_verem: a team of researchers just proved you don't need a bigger model, you need a smarter plan researchers from Tsinghua and …
Researchers from Tsinghua and South China University of Technology introduced Atomic Task Graph (ATG), a framework that enables 7B-8B open-source models to surpass GPT-4 on complex agent benchmarks without fine-tuning, by using directed graph-based planning and internal simulation to drastically reduce hallucination rates.
Training Reasoning Models to Generate Diverse Scientific Ideas
Researchers propose a set-level reinforcement learning objective that trains language models to generate diverse, high-quality, document-grounded scientific ideas, with an 8B model outperforming Claude Opus 4.8 and GPT-6 Astra on creative utility.
@yuetai12575: Excited to see the gains further survive as model scaling!
The article reports initial results from OpenRSI's Marin-Scaling-Ladder experiment, where AI agents autonomously propose, implement, and evaluate new optimizers across model scales from 550M to 2.5B, discovering PSPR with Codex (GPT-5.6).
DeepReinforce releases Ornith-1.0 open-source coding models (2 minute read)
DeepReinforce open-sources Ornith-1.0, a family of self-improving coding models from 9B to 397B parameters, trained on Gemma 4 and Qwen 3.5 foundations, featuring a novel RL approach that learns to generate its own scaffolds.
Agents-A1-4B (Qwen3.7-4B ???) : Scaling the Horizon, Not the Parameters
Agents-A1-4B, a compact 4B-parameter model from InternScience, achieves state-of-the-art results across multiple agentic and reasoning benchmarks, outperforming larger models by scaling capabilities rather than parameters.