Check out the full details below. Everything is versioned so existing servers will still work just fine. SDK updates are on the way. https://
github.com/modelcontextpr
otocol/specification/tree/main/docs/specification/2025-03-26
…
OPEN SOURCE
-
Model Context Protocol Specification Update Released March 2025
By
–
-
New MCP Spec Version Released with Major Updates
By
–
A new version of the MCP spec was finalized today. Some of the major changes:
– Auth framework based on OAuth 2.1
– Replaced the previous HTTP+SSE transport with Streamable HTTP transport
– Support for JSON-RPC batching
– Tool annotations for better describing tool behavior -
ByteDance releases InfiniteYou AI portrait generator
By
–
TikTok parent ByteDance released InfiniteYou, an open-source AI portrait generator
— Rowan Cheung (@rowancheung) 26 mars 2025
It requires you to upload just one photo and produces consistent portraits in different styles, in just a few secondspic.twitter.com/TFIf1078UeTikTok parent ByteDance released InfiniteYou, an open-source AI portrait generator It requires you to upload just one photo and produces consistent portraits in different styles, in just a few seconds
-
1 billion open source methods with LoRAs
By
–
You have 1 billion methods to do that in open source with LoRAs. Gone are the days when OpenAI released features never seen before.
-
Accessing Legacy AI Model Versions for Content Generation
By
–
You can still generate this kind of stuff by going to the old versions of MidJourney or Stable Diffusion. 🙂 They don't need to be lost forever!
-
Signal Claims Gold Standard in Encrypted Private Communications
By
–
I wouldn’t say that Will and I are battling but I do disagree. Because there are big differences between Signal and WhatsApp. Signal is the gold standard in private comms. We’re open source, nonprofit, and we develop and apply e2ee and privacy preserving tech across our system
-
Alibaba Qwen2.5-VL-32B-Instruct: Advanced Vision-Language Model
By
–
The Qwen team of Alibaba open-sourced Qwen2.5-VL-32B-Instruct, a new vision-language model —Features enhanced mathematical reasoning and visual capabilities
—Beats Qwen2-VL-72B, Mistral Small 3.1-24B, and GPT-4o-0513
—Apache 2.0 licensed -
DeepSeek V3-0324: Open-Source AI Model with Enhanced Math Capabilities
By
–
DeepSeek released V3-0324, an updated version of its V3 non-reasoning AI
— Rowan Cheung (@rowancheung) 25 mars 2025
—A 641GB model but capable of running on high-end PCs
—Uses MoE, activating only 37B params/token and reducing compute
—Upgraded Math and Coding capabilities
—Open-source under MITpic.twitter.com/sskVI1tIFKDeepSeek released V3-0324, an updated version of its V3 non-reasoning AI —A 641GB model but capable of running on high-end PCs
—Uses MoE, activating only 37B params/token and reducing compute
—Upgraded Math and Coding capabilities
—Open-source under MIT -
DeepSeek V3 Latest Version Improves Math Coding Reasoning
By
–
The latest version of DeepSeek V3 is now available! Today’s update improves its math, coding, and general reasoning performance, making it one of the most powerful open-source LLMs available. (1/2)
-
Together AI launches Together Chat with open source models
By
–
Together AI released Together Chat with a bunch of open source models including DeepSeek R1 and Llama 3.3 70B. It also supports Thinking, Web Search and chat branching.