TestingCatalog AI News πŸ—ž
4.64K subscribers
2.87K photos
372 videos
40 files
3.85K links
Reporting AI nonsense. A future news media, driven by virtual assistants πŸ€–
Download Telegram
StepFun launched Step 3.5 Flash open-source model

Step 3.5 Flash is a high-speed, cost-efficient language model built for private local use, deploying only 11B of 196B parameters per token. It supports 256K token contexts, runs on consumer hardware, and minimizes cloud reliance.

πŸ—ž #ai
❀3πŸ‘1
OpenAI partnered with Snowflake to integrate AI solutions from OpenAI into Cortex AI and Snowflake Intelligence.

"OpenAI and Snowflake have entered a multi-year, $200 million partnership that brings OpenAI frontier intelligence directly into Snowflake⁠."
πŸ‘2
Z AI may drop GLM-5 in February πŸ‘€
8❀53
This media is not supported in your browser
VIEW IN TELEGRAM
BREAKING 🚨: OpenAI launched a dedicated Vibe coding Codex app for macOS. Codex also supports Skills and automations!

Everyone is a developer now πŸ‘€
❀8🀩4πŸ‘¨β€πŸ’»3
Claude mobile app will get a new tool notifications toggle soon and support Plan Mode in Claude Code.

Tools will arrive on mobile too πŸ”₯
πŸ‘33
Anthropic is about to drop Sonnet 5 during Super Bowl week

Leaks reveal Anthropic’s Claude Sonnet 5 may debut around this week, showing notable performance in math and coding, strong visual code generation, targeting cost-efficient use for developers and general users.

h/t @vvirtr

πŸ—ž #claude
❀5πŸ‘42
xAI joined SpaceX πŸ‘€
πŸ₯±1151
Perplexity Finance now has a heatmap widget πŸ”₯
❀10πŸ‘54
Now it is possible to specify the model used in scheduled tasks for Manus AI.

Now I need more tasks πŸ‘€
πŸ‘5
OpenAI launches Codex app for Mac to manage coding agents

OpenAI released the Codex app for macOS, allowing users to manage multiple coding agents for long tasks with isolated workspaces, reusable workflows, and automation. It includes secure sandboxing and supports Git worktrees.

πŸ—ž #chatgpt
πŸ‘3😍1
A new 72% acheivement submission for ARC-AGI-2. So far, it is the second multi-model system that outperformed single-model solutions.

"It runs the same task through GPT-5.2, Gemini-3, and Claude Opus 4.5 in parallel."

We need new benchmarks πŸ‘€
❀2
NotebookLM mobile apps now support Video Overview generation
❀9πŸ”₯3
Which models is that? πŸ‘€

β€œCyberpunk hacker robot working in front of many monitors”
πŸ‘€3