@AdinaYakup: BigBang-v1a self-evolving LLM from endless frontier lab in Shanghai - Self-evolving training with AI generated frontier…
Summary
BigBang-v1 is a self-evolving 36B LLM from Endless Frontier Lab in Shanghai, trained with AI-generated frontier tasks and achieving strong performance with only 10K high-quality examples across science, coding, tool use, and long context.
View Cached Full Text
Cached at: 08/08/26, 05:00 AM
BigBang-v1🔬a self-evolving LLM from endless frontier lab in Shanghai
- Self-evolving training with AI generated frontier tasks
- 36B model matching much larger models
- Strong performance with 10K high-quality examples
- Built for science, coding, tool use, and long context https://t.co/ZUaILV2ZVQ
Similar Articles
@AdinaYakup: Intern S2 preview A scientific multimodal model from Shanghai AI Lab @intern_lm 35B matches their own 1T model on scien…
Shanghai AI Lab releases Intern S2, a 35B scientific multimodal model that matches their own 1T model on science benchmarks, introducing Task Scaling as a new scaling dimension. Licensed under Apache 2.0.
Training LLM Agents for Spontaneous, Reward-Free Self-Evolution via World Knowledge Exploration
This paper proposes a method to train LLM agents with intrinsic meta-evolution capabilities, enabling spontaneous self-improvement without external rewards at inference time. Applied to Qwen3-30B and Seed-OSS-36B, the approach yields a 20% performance boost on web navigation benchmarks, with a 14B model outperforming Gemini-2.5-Flash.
@KaiZhang_CS: Check out one of the best open-source search agents trained by @jianxie_ !! glad to see early experience methods work o…
Yu Su's team trained a frontier Deep Research Agent on an academic budget using 8K synthetic samples and RL, releasing fully open training infrastructure and models from 2B to 35B parameters.
LiquidAI/LFM2.5-230M
Liquid AI released LFM2.5-230M, a compact 230M-parameter hybrid model optimized for on-device deployment with fast edge inference speeds (213 tok/s on Galaxy S25 Ultra) and built for agentic tasks via reinforcement learning.
Developing open source LLM from ground up from pretrain - rlhf(PPO/GRPO)
A developer shares progress on training a 7B parameter open source LLM from scratch using a DeepSeek architecture optimized for low VRAM, with the goal of democratizing AI development and eventually surpassing large proprietary models.