Moonshot has launched Kimi K3, a 2.8-trillion-parameter multimodal model featuring a 1-million-token context window and optimizations for faster long-context decoding and agentic coding. Google reportedly delayed its Gemini 3.5 Pro model due to performance issues, particularly in coding, though the company states it remains in partner testing. Additionally, a new harness called Schema has been developed to improve frontier models' performance on benchmarks like ARC-AGI-3 by enabling them to write, test, and execute programs based on game mechanics. AI
IMPACT New large-context models like Kimi K3 may accelerate agentic workflows and complex coding tasks.
RANK_REASON Cluster includes a new model release (Kimi K3) from a frontier lab (Moonshot). [lever_c_demoted from frontier_release: ic=1 ai=1.0]
- ARC-AGI-3
- Claude Code
- Gemini 3.5 Pro
- GPT-5.6
- Kimi K3
- LM Studio Bionic
- Moonshot
- NVIDIA Nemotron 3 Embed
- Schema
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
