Kimi K3 Released: Nearly 3-Trillion-Parameter Open Model Rivals Closed-Source Frontier
Aizkmusic · x · 2026-07-30
Moonshot has released Kimi K3, now available on Featherless. Scaling to 2.8 trillion total parameters with roughly 104 billion active per token, it is the first open-weight model approaching the 3-trillion parameter mark.
Key Upgrades & Benchmarks:
- Architecture: Utilizes 16 out of 896 experts and introduces Kimi Delta Attention, offering 2.5× better scaling efficiency than K2.
- Multimodality: Upgrades native vision capabilities with the new MoonViT-V2 vision encoder.
- Performance: Scores 88.3 on Terminal-Bench 2.1 and 93.5 on GPQA Diamond, running toe-to-toe with top closed-weight models like Claude and GPT.
More from Models
- Reflection launches Beam 1.0, its first open-weight model built for coding, reasoning and agentic tasks — tomaarsen · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- JevBench v1.7.46 leaderboard: new d1-ultra tops 32 API systems at 76.1 composite — airesearch12 · 2026-10-16
- LMArena launches Agent Arena and Code Arena leaderboards for model comparison — arena · 2026-10-16
- Third-party tests find NVIDIA Nemotron-3.5 Bolt underperforms nano on pandas/numpy and decision-making tasks — norenEnmotalen · 2026-10-16
- Reflection open-sources Beam under Apache 2.0 with 40+ partner ecosystem — bhutanisanyam1 · 2026-10-16