We've evaluated a lot of base models on perplexity-based evals and Kimi k2.5 proved to be the strong...
TL;DR · AI 摘要
Cursor团队称Kimi k2.5在困惑度评估中表现最佳，经持续预训练和高算力强化学习后推出Composer-2模型。
核心要点
- Kimi k2.5在多项基础模型评估中表现最优
- Composer-2结合了CPT、高算力RL和Fireworks推理采样技术
- 团队承认此前博客未提及Kimi基础模型是疏漏
After that, we do continued pre-training and high-compute RL (a 4x scale-up).
The combination of the strong base, CPT and RL, and Fireworks' inference and RL samplers make" / X
Aman Sanger on X: "We've evaluated a lot of base models on perplexity-based evals and Kimi k2.5 proved to be the strongest! After that, we do continued pre-training and high-compute RL (a 4x scale-up). The combination of the strong base, CPT and RL, and Fireworks' inference and RL samplers make" / X
Don’t miss what’s happening
Aman Sanger 
We've evaluated a lot of base models on perplexity-based evals and Kimi k2.5 proved to be the strongest! After that, we do continued pre-training and high-compute RL (a 4x scale-up). The combination of the strong base, CPT and RL, and Fireworks' inference and RL samplers make Composer-2 frontier level. It was a miss to not mention the Kimi base in our blog from the start. We'll fix that for the next model.
Quote
Kimi.ai
@Kimi_Moonshot
·
Mar 20
Congrats to the @cursor_ai team on the launch of Composer 2! We are proud to see Kimi-k2.5 provide the foundation. Seeing our model integrated effectively through Cursor's continued pretraining & high-compute RL training is the open model ecosystem we love to support.
·
150
210
2.5K
446
Read 150 replies