DeepSeek published an updated R1 model on HuggingFace. Seems like this updated model is already powering DeepSeek apps as users spotted longer thinking times and performance improvements.
LLMS
-
Claude 2.5 Pro to Close Capability Gaps Soon
By
–
new 2.5 pro should close the gaps in a couple of weeks
-
Excitement for Claude Launch in Factory Platform
By
–
congrats on the launch! excited to see what folks can do with claude in Factory
-
Evals and Benchmarks Don’t Capture What You Think
By
–
There’s also so many layers to most to really understand what they are capturing, it’s pretty interesting. Most evals and benchmarks do not capture what you would initially assume.
-
Anthropic Releases Go and Ruby SDKs for Developers
By
–
Today we made the Go and Ruby Anthropic client SDKs generally available! We now offer SDKs for Python, TypeScript, Java, Go, and Ruby.
-
AI Models’ Reasoning Capabilities and High-Stakes Limitations
By
–
AI isn’t ready for high-stakes contexts where precision is critical. In this year’s AI Index, we chart the progress of top models’ complex reasoning capabilities. See more insights from the technical chapter of the #AIIndex2025: https://
stanford.io/43w3rEo -
Gemini Models and App: Past, Present, and Future
By
–
A conversation with @joshwoodward and @tulseedoshi (two of my favorite people) about the past, present, and future of Gemini models + app pic.twitter.com/Bn8lTFMtuJ
— Logan Kilpatrick (@OfficialLoganK) 28 mai 2025A conversation with @joshwoodward and @tulseedoshi (two of my favorite people) about the past, present, and future of Gemini models + app
-
AI Reasoning Toggle Feature Coming June 2026
By
–
The ability to turn off reasoning is coming in June! The team is working hard, and sadly is not as simple as a 2 step prompt : )
-
Codestral Embed: New Code Embedder Model Released
By
–
Very happy to release Codestral Embed, our first code embedder. It can use up to 3072 dimensions, ordered by relevance. Using embeddings of 256 dimensions with int8 precision is already sufficient to outperform all existing models.