Leaks & rumours
Report Says Fable 5.5 Is Rolling Out to Enterprise Clients
A post on X claims Fable 5.5 has rolled out to enterprise customers including JP Morgan and Meta
Prompt Blueprints / Topics
The latest in AI.
Releases, research and the stories shaping AI. Reporting across the companies and models that matter.
More stories
AI releases
Nanbeige4.3-33B Open-Model Agentic Lead at 30B Scale
Nanbeige has released the open-weight Nanbeige4.3-33B, claiming leading agentic coding and cowork scores among ~30B open models based on its own benchmark…
News
Microsoft's EF Core Repo Reportedly Names GPT-6.1 Terra as Judge
A post by @chetaslua says Microsoft's EF Core team set an unreleased-looking GPT-6.1 Terra as the default judge model in 18 eval configs.
News
Over 1,000 Mathematicians Join the Association for Human Mathematics
More than 1,000 mathematicians have joined the Association for Human Mathematics, and 654 have also joined an optional AI-free caucus that bars AI use in…
News
Claude Code's @trq212 Says He Has Left Anthropic
Thariq Shihipar (@trq212), a Claude Code engineer, says in his own X post that he has left Anthropic after over a year. No Anthropic statement was found, and…
Leaks & rumours
Gemini 'createDecision' Leak: Google Already Has a Decision Tool
An X post claims Google has a Decision model and a Gemini API createDecision method, but the Gemini API release notes through October 13 show neither. Google…
Leaks & rumours
Stealth Image Models 'shooting-star' Spotted in Arena Scout Alerts
Two unnamed image models, shooting-star-green-alpha and shooting-star-red-alpha, appeared in Arena Scout alerts on Oct 23, 2026. A poster calls them new GPT…
AI releases
MiniMax M3.1 Flash Ships Open Weights at a Quarter of M3's Price
MiniMax has released M3.1 Flash with open weights on Hugging Face and API pricing of $0.15 per million input tokens, a quarter of M3's list rate. MiniMax's own…
Leaks & rumours
Gemini 4 Carbon Spotted in Model List Beside Gemini 4 Flash
A screenshot from @bidyut_io shows gemini-4-carbon-preview below gemini-4-flash-preview. Google's docs list only Gemini 4 Argon, and neither unlisted name has…
AI releases
Canva's First Image Model Lands Behind Only OpenAI
Canva's first in-house image model, Craft, is live in Dream Lab for all users and on the public API. Canva's chart puts it 4th on DesignArena at 1340 Elo…
News
Banks Find Unvetted AI Agents Inside Customer Accounts
Axios reports, citing unnamed bank and industry executives, that major U.S. banks are detecting unauthorized third-party AI agents in customer accounts. The…
AI releases
Grok Bot can now build reports and dashboards that stay updated
xAI's @bot account says Grok Bot can now build editable pages that it keeps up to date. The docs, last updated before the announcement, don't describe the…
News
Railway already reset some secrets; rotate your AI keys now
Railway says an unauthorized party may have decrypted some customers' environment variables, and it tells every user to rotate stored secrets, including AI…
News
OpenAI says a plan-change error cut Codex allowances on Oct 16, and is adding $500 in credits
An OpenAI Support email, shared on X, says an error kept Pro 200 allowances in place after downgrades. The correction cut weekly allowances abruptly, and…
AI releases
Ornith-1.5-122B Released: MIT-Licensed Open Weights, Quantised Builds and Ollama Support
Ornith-1.5-122B is a new open-weight reasoning MoE model under the MIT license with a 262,144-token context. This article covers how to run it with vLLM,…
News
OpenAI Platform Adds a Decisions Playground for Testing the Decisions API
OpenAI's Platform sidebar now shows a Decisions item marked "New" that opens a playground for the Decisions API, per a post on X. OpenAI's guide says…
AI releases
Mercury 3: Haiku-Class Quality at About 4x the Speed, Per Inception's Own Numbers
Inception says its diffusion LLM Mercury 3 matches or beats Claude Haiku 5.5 and GPT-6 Luna on quality while running 4–8x faster. Here is the pricing worked…
Leaks & rumours
GLM-Omni Is a Z.ai Research Model in a New Paper, Not a Confirmed GLM-5.5
A leak post claims Z.ai's GLM-Omni is probably GLM-5.5. The paper names the model but describes a research build on a GLM-4.1V-Thinking base, and Z.ai's docs…