The OpenAI Agents SDK empowers you to build lightweight and user-friendly agentic AI applications with minimal abstractions. It’s a production-ready upgrade of our previous agent experimentation, Swarm. The SDK includes primitives like…
Tag: ai
Mistral OCR, MCP – Model Context Protocol, Windsurf
Mistral OCR, a new Optical Character Recognition API, unlocks the collective intelligence of digitized information by accurately understanding documents. The API, available on the developer suite la Plateforme, is ideal for multimodal…
Ollama MLX support + LM Studio comparison
Soon (hopefully) Mac user will get MLX support for Ollama: https://github.com/ollama/ollama/pull/9118 – it’s said to be ~40% speed improvement, but it’s not for every model (it has to be MLX version). For…
Cursor how helpful is it?
About Cursor, I know from a long, long time, but the first time I decided to give it a try. What are my thoughts on it:– Superb when to run infrastructure of…
King is back
Yes, it’s about new Claude 3.7 Sonnet … I even don’t have time to check Grok 3 when I’m back to Claude… Claude 3.7 Sonnet is the first hybrid reasoning model and…
PLLuM another Polish model
PLLuM is a family of large language models (LLMs) specialized in Polish and other Slavic/Baltic languages, with additional English data incorporated for broader generalization. Developed through an extensive collaboration with various data…
Zyphra/Zonos
Zonos-v0.1 is open-weight text-to-speech model, is trending on HF. It was trained on over 200k hours of diverse multilingual speech, resulting in expressiveness and quality comparable to — or even surpassing —…
Lumina Image 2.0
Lumina 2.0 is a solid base model that fulfills our expectations for SD3/3.5. It’s truly open-source with an Apache 2.0 license, which is a significant advantage. In terms of comparison with Flux,…
DeepSeek R1 – model quantization for customer grade cards
This model has shaken the NASDAQ, so it is not a big surprise that people started to adapt it further. The requirements are very high for home / hobby usage, but already…
DeepSeek R1
DeepSeek R1 is available for regular users, in examples it beats o1 reasoning, there are also smaller models available 1.5B so quite a good potential to be tested. Exolabs of course tested…