Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Kimi k1.5: A Next Generation Multi-Modal LLM Trained with Reinforcement Learning on Advancing AI with Scalable Multimodal Reasoning and Benchmark Excellence https://t.co/z4vGN4QL40 #ReinforcementLearning #LargeLanguageModels #KimiK1dot5 #AIInnovation #MultimodalAI #ai #news #… https://t.co/nAV934F6DS
Kimi k1.5: A Next Generation Multi-Modal LLM Trained with Reinforcement Learning on Advancing AI with Scalable Multimodal Reasoning and Benchmark Excellence Researchers from the Kimi Team have introduced Kimi k1.5, a next-generation multimodal LLM designed to overcome these… https://t.co/N7LgUfhXQs
Kimi k1.5: Multimodal AI model using RL and long/short CoT reasoning https://t.co/fU9fwZ4ZRo
The Kimi Team has unveiled Kimi k1.5, a next-generation multimodal AI model that integrates reinforcement learning with advanced reasoning capabilities. This model is designed to excel in various benchmarks, showcasing state-of-the-art (SOTA) performance in short-chain of thought (short-CoT) tasks, outperforming competitors such as GPT-4o and Claude 3.5 by over 550% on the AIME and LiveCodeBench benchmarks. Additionally, Kimi k1.5 demonstrates comparable performance to its predecessors on long-chain of thought (long-CoT) tasks across platforms like MathVista and Codeforces. The model supports a long context of 128k tokens, enhancing its usability for complex applications.
