AI Leaks & Rumors 👀 - TestingCatalog (Page 2)

29 Dec 2025 · 2 min read

Notion tests new internal models and AI-first Workspaces

Notion is testing a redesigned AI-first workspace with a new sidebar, toggles for AI-powered features, and an AI credit usage tracker.

27 Dec 2025 · 1 min read

Mistral AI working on Workflow builder for its platform

Mistral is testing a Workflow Builder and Connectors feature, as well as streamlining the Le Chat attachments menu.

23 Dec 2025 · 2 min read

Exclusive: Google tests 30-minute audio Lectures on NotebookLM

Google NotebookLM is testing a new “Lecture” Audio Overview format with language selection support and upcoming British English narration.

20 Dec 2025 · 1 min read

Meta AI to get Memories and Custom Prompts soon

Meta AI is testing memories and custom prompts, found hidden in settings, aiming to bridge the gap with top-tier AI labs.

19 Dec 2025 · 2 min read

Luma AI working on Realms, Ray 3 Flash and Image V2

Luma Labs is testing new AI models like Ray 3 Flash and an Image V2 option, with hints at a potential World Model.

19 Dec 2025 · 2 min read

Anthropic working on Planning Mode for Claude on Chrome

Anthropic’s Claude for Chrome extension now reaches all paid users, adding shortcut templates and a planning mode in the future.

18 Dec 2025 · 1 min read

xAI works on Imagine API and Grok API Playground

xAI opens new opportunities for developers with the Voice Agent API and is developing new Playground and Imagine API internally.

18 Dec 2025 · 2 min read

Exclusive: Early look at Claude Tasks Mode from Anthropic

Anthropic’s Claude task mode will offer step-by-step task execution, artefact management, and context file control in a streamlined dual-panel UI

18 Dec 2025 · 1 min read

OpenAI readies Codex Caribou model upgrade with GPT-5.2 base

OpenAI is preparing a Codex model upgrade, codenamed Caribou, with GPT-5.2 integration to advance coding tools for developers on Codex.

16 Dec 2025 · 2 min read

Anthropic preparing new Agentic Tasks Mode for Claude

Anthropic testing Claude's Agent mode with a new interface for tasks, to introduce new modes for research, analysis, writing, and building.

11 Dec 2025 · 1 min read

Google tests Gemini 3 Flash model on LM Arena

Google is testing Fierce Falcon and Ghost Falcon models on LM Arena, signalling a potential Gemini 3 Flash or Gemini 3 Pro GA launch.

11 Dec 2025 · 1 min read

OpenAI debuts Garlic model on LM Arena for public testing

OpenAI’s new Robin High model is now on LM Arena for benchmarking, outperforming previous contenders and offering a glimpse at next-gen AI progress.

10 Dec 2025 · 2 min read

OpenAI testing new Image-2 models on LM Arena

OpenAI is set to debut new image generation models, Image-2 and Image-2-mini, promising higher detail and colour accuracy for ChatGPT and creative tools.

10 Dec 2025 · 2 min read

Notion tests unreleased GPT-5.2 model as OpenAI launch nears

OpenAI's GPT-5.2 model is rumored for release soon, with leaks from Notion hinting at early testing and speculation around a potential launch date.

8 Dec 2025 · 2 min read

Anthropic works on Agent Mode for Claude and new avatars

Anthropic is testing a dedicated agent mode for Claude, with a home screen toggle and structured task flows, plus a new pixel avatar feature in settings.

7 Dec 2025 · 2 min read

Exclusive: Google set to launch Gemini Nano Banana 2 Flash

Google is preparing to launch Nano Banana 2 Flash within Gemini, offering similar performance to the Pro model at lower costs for broader accessibility.

6 Dec 2025 · 1 min read

Google prepares Gemini Live with screen sharing for desktop web

Google is testing Gemini Live support for desktop web, with features like screen sharing and live translation previously limited to mobile devices.

6 Dec 2025 · 2 min read

Microsoft prepares GPT-5.1, Reminders and Projects for Copilot

Microsoft Copilot is testing GPT-5.1 in Smart mode, Reminders, and Projects to boost task and chat management.