High scores on AIME! The new DeepSeek-R1-Lite-Preview is here.
DeepSeek-R1-Lite-Preview is now live: unleashing supercharged reasoning power!
Original: 🚀 DeepSeek-R1-Lite-Preview is now live: unleashing supercharged reasoning power! | DeepSeek API Docs
Importance: 新機能として多くのユーザーに影響を与えるため
Summary
DeepSeek has launched the new DeepSeek-R1-Lite-Preview. This model demonstrates high performance on AIME and MATH benchmarks, offering a transparent thought process in real-time. Open-source models and API will be available soon.
Key Points
- High performance on AIME and MATH benchmarks
- Transparent thought process in real-time
- Open-source models and API coming soon
- Score improves with increased reasoning length
View developer notes (APIs, breaking changes, migration)
DeepSeek-R1-Lite-Preview achieves o1-preview-level performance on AIME and MATH benchmarks. It shows steady score improvements as reasoning length increases, with open-source models and API set to be released soon.
Source: https://api-docs.deepseek.com/news/news1120
Outlet: DeepSeek Release Notes
This article is an AI-generated summary (OpenAI GPT-4o-mini) of publicly available information from Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, Sakana, and other vendors. The original source URL is always provided in accordance with fair-use citation requirements. Summaries are AI-generated and may contain mistranslations or misinterpretations. Always verify details with the original source.