MoonshotAI: Kimi K2 0905 (exacto)

MoonshotAI Kimi K2 0905 (exacto) is the September update of the Kimi K2 series, a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI. Featuring 1 trillion total parameters with 32 billion active per forward pass, this model significantly enhances agentic coding with higher accuracy and better generalization. It also improves frontend coding, delivering more aesthetic and functional outputs for web, 3D, and related tasks. Kimi K2 is optimized for advanced tool use, reasoning, and code synthesis. This model boasts an impressive context window of 262K tokens and a maximum output of 4K tokens, making it ideal for complex, long-form tasks. It excels across various benchmarks, including coding (LiveCodeBench, SWE-bench), reasoning (ZebraLogic, GPQA), and tool-use (Tau2, AceBench). Access this PRO-tier model for $0.60/2.50 per 1M tokens (input/output) and leverage its functions, code generation, and streaming capabilities for your projects.