This media is not supported in your browser
VIEW IN TELEGRAM
Microsoft announced that new Maia 200 AI accelerators are becoming available on Azure for advanced AI workloads.
"10+ PFLOPS FP4 throughput, ~5 PFLOPS FP8, and 216GB HBM3e with 7TB/s of memory bandwidth"
"10+ PFLOPS FP4 throughput, ~5 PFLOPS FP8, and 216GB HBM3e with 7TB/s of memory bandwidth"
❤4🔥3
Dario Amodei explicitly argues there is “a strong chance” that “powerful AI”, smarter than Nobel laureates across domains and able to operate as a “country of geniuses in a datacenter” with millions of instances, arrives within the next few years, possibly as soon as 2027.
He notes that Claude Sonnet 4.5 could recognise when it was being evaluated and adjust behaviour, and that when researchers altered a model’s “beliefs” to make it think it was not being evaluated.
He notes that Claude Sonnet 4.5 could recognise when it was being evaluated and adjust behaviour, and that when researchers altered a model’s “beliefs” to make it think it was not being evaluated.
🔥5👍2🤡1
This media is not supported in your browser
VIEW IN TELEGRAM
Anthropic released interactive apps for Claude. Interactive apps can respond with interactive UI Widgets to enable additional usecases.
"MCP Apps is a new extension to MCP that lets any MCP server deliver an interactive interface within any supporting AI product"
"MCP Apps is a new extension to MCP that lets any MCP server deliver an interactive interface within any supporting AI product"
🔥7❤6👍4
Qwen3-Max-Thinking debuts with focus on hard math, code
Qwen3-Max-Thinking is Alibaba Cloud’s advanced reasoning model for math, coding, and agent workflows, now accessible via Qwen Chat and Model Studio. It supports long-context tasks, tool use, and selective compute for accuracy-critical prompts.
🗞 #ai
Qwen3-Max-Thinking is Alibaba Cloud’s advanced reasoning model for math, coding, and agent workflows, now accessible via Qwen Chat and Model Studio. It supports long-context tasks, tool use, and selective compute for accuracy-critical prompts.
🗞 #ai
TestingCatalog
Qwen3-Max-Thinking debuts with focus on hard math, code
Qwen3-Max-Thinking, Alibaba Cloud's new flagship reasoning model, is now in Qwen Chat and Model Studio, targeting tough math, code, and agent workflows.
👍3❤1
Anthropic integrates interactive MCP apps into Claude
Anthropic updates Claude to support direct use of tools like Asana, Slack, Figma, and Box within its interface. Users on paid plans can now collaborate on projects without switching apps, using real-time UIs powered by the open-source MCP protocol.
🗞 #claude
Anthropic updates Claude to support direct use of tools like Asana, Slack, Figma, and Box within its interface. Users on paid plans can now collaborate on projects without switching apps, using real-time UIs powered by the open-source MCP protocol.
🗞 #claude
TestingCatalog
Anthropic integrates interactive MCP apps into Claude
What's new? Anthropics updated Claude to let users access Asana, Slack, Figma and Box tools in chat; developers can build apps with MCP and view live tool content mid-chat;
👍3
Anthropic is working on a new inline voice mode UI for its mobile apps. Users will be able to seamlessly switch between text and voice conversations.
👍5
TestingCatalog AI News 🗞
Anthropic is working on a new inline voice mode UI for its mobile apps. Users will be able to seamlessly switch between text and voice conversations.
Besides this 👀
- Claude Code is about to get prompt suggestions.
- A new Thinking effort selector will be added to the model selector.
- Claude Code is about to get prompt suggestions.
- A new Thinking effort selector will be added to the model selector.
👍5
This media is not supported in your browser
VIEW IN TELEGRAM
BREAKING 🚨: Kimi K2.5 open-source model is now live on Kimi Chat and APIs with a leading 50% score on HLE benchmark!
It comes along with an Agentic Swarm feature, where up to 100 sub-agents would be working on a problem in parallel (Available in beta for some customers)
It comes along with an Agentic Swarm feature, where up to 100 sub-agents would be working on a problem in parallel (Available in beta for some customers)
🔥5👍2
TestingCatalog AI News 🗞
BREAKING 🚨: Kimi K2.5 open-source model is now live on Kimi Chat and APIs with a leading 50% score on HLE benchmark! It comes along with an Agentic Swarm feature, where up to 100 sub-agents would be working on a problem in parallel (Available in beta for…
Benchmarks 👀
❤🔥4❤3
TestingCatalog AI News 🗞
It turned out, in fact, that Clawdbot is all you need. This is the best thing you can test at this moment. Have you tried it yet? 🦀
ICYMI: A popular GitHub project Clawdbot is now a Moltbot as Anthropic pushed the project to get a new name.
Molty 🦞
Molty 🦞
😭6👍1
This media is not supported in your browser
VIEW IN TELEGRAM
Mistral AI released Mistral Vibe 2.0, an upgraded SWE agent CLI with subagents, skills support and new unified agent modes. Now available on Team and Pro plans.
Terminal Testing Time 👀
Terminal Testing Time 👀
🔥4👍3
This media is not supported in your browser
VIEW IN TELEGRAM
Manus AI now supports Skills, a common standard introduced by Anthropic, so now you can reuse them in Manus as well.
Skill is not an issue anymore 👀
Skill is not an issue anymore 👀
🔥3👍2
TestingCatalog AI News 🗞
Anthropic works on customizable Commands for Claude Code Anthropic is developing a Customize section for Claude, consolidating tools like Skills, Connectors, and a new Commands feature to support tailored workflows and modular use, aiming to serve professional…
Seems like Anthropic is preparing to release a dedicated Plugins section for Connectors and Skills, as it now appears on Claude Desktop. However, it is not clickable yet.
This feature was named "Customize" earlier, during the development phase.
The customizable commands option was removed after this publication, and yet unclear if those will be discontinued (very likely).
This feature was named "Customize" earlier, during the development phase.
The customizable commands option was removed after this publication, and yet unclear if those will be discontinued (very likely).
🔥5👍1