Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
slimethudm.github.io · Low evidenceSlime is an LLM post-training framework for reinforcement learning scaling, offering high-performance training by connecting Megatron with SGLang and flexible data generation through custom interfaces. It powers models like GLM 4.7, GLM 4.6, and GLM 4.5, and supports Qwen3, DeepSeek V3, DeepSeek R1, and Llama 3 series.
GitHubslimeParse Score