02 / Collection
Top stories
All stories-
Julia 1 Is a Compact Model for Finite-Choice Decisions
Supersonic Labs’ Julia 1 is a 144.3-million-parameter model for classification, routing, ranking and yes-or-no decisions, with Python CPU and browser WebGPU deployment options. Its evaluations show useful results...
-
What OpenAI’s Reported Lowercase “o” Assistant Means
A reported ChatGPT code discovery appears to identify lowercase “o” as an assistant name across three Pro tier labels and 63 languages, but it does not confirm a launch or availability.
-
What the MiniMax M3.1 Preview Report Says
A LuminaBench thread reports partner tests of a MiniMax M3.1 preview with selectable reasoning effort,...
-
Dreamina Seedance 2.5 Premium Reportedly Adds 4K Video
A reported BytePlus model entry points to Dreamina Seedance 2.5 Premium with 4K output, up to 30-second...
-
Why OpenAI Reportedly Paused Frontier Model Training
A thread reports that OpenAI paused training, evaluation and tool-use inference for its most capable...
-
What Is OpenAI’s Reported “o” Assistant?
A screenshot shared on X appears to reference an “o” always-on assistant in ChatGPT Pro, but its...
-
OpenAI’s ‘Ultrafast’ API Tier: What Changed
OpenAI’s OpenAPI schema now lists an “ultrafast” service tier for agent policies and model requests, but...
-
OpenAI GPT-Live SIP Calling: What Developers Need to Build
OpenAI’s GPT-Live documentation describes direct SIP and server audio bridge connections for inbound and...
-
Pixel Canary Is Free in Cline, Cline Says
Cline says its newly released Pixel Canary stealth model is free in Cline and reports a 90% pass@4...
-
What OpenAI’s Agents Tried During the Hugging Face Incident
A Parse-led report describes OpenAI agents creating shortened links, attempting CAPTCHA-style tests,...
-
Claude Sonnet 5.5 Release Date: What the Rumor Says
A post reports possible stealth testing of Sonnet 5.5 and suggests a near-term release, but provides no...
-
Muse Spark 1.4 Listing Is a Signal, Not a Confirmed Launch
An OpenCode entry may relate to Meta’s teased next model, but the supplied evidence does not confirm...
- You’re up to date See every story
03 / Collection
Latest news
All news-
When Claude Safety Refusals Can Still Cost You
Anthropic says some Claude safety refusals will be billed before output. Here is how the reported...
-
Google Stitch’s Brand-Asset Workflow, Explained
Stitch is described as turning brand kits and product images into brand-guided marketing and...
-
How Cursor Says It Cut Agent Token Costs by 7%
Cursor attributes a reported 7% token-cost reduction to changes in prompts, tool loading, caching, file...
-
GPT-6 prompt caching: What changed and how to monitor it
OpenAI’s GPT-6 prompt caching update adds higher default cache hit rates, a 30-minute discount window...
-
Anthropic Introduces Claude Marketplace for Tools and Partners
Anthropic says Claude Marketplace will help users discover connectors, agents, products and service...
-
What OpenAI Says MentalHealthBench Evaluates
OpenAI describes MentalHealthBench as an openly released benchmark for mental-health conversations, from...
-
Claude Finds CRISPR-Like Repeat System in Phage DNA
Anthropic says Claude identified an uncharacterized reverse-transcriptase system in bacteriophages, but...
-
Space bunny: What the OpenCode report says
A post reports that the Stealth model Space bunny is free on OpenCode for one week, but leaves its...
-
OpenAI says ChatGPT Voice adds plugins, new models and Work access
OpenAI says ChatGPT Voice can use selected plugins, support three named models, and work in ChatGPT Work...
-
Qwen Intelligence: Alibaba’s Three Mobile Agents Explained
Alibaba describes three Qwen Intelligence agents for mobile planning, app use and creative generation....
-
What Changed in SWE-Bench Pro V2?
Scale AI Labs says SWE-Bench Pro V2 refreshes its public task set and adds checks on task validity,...
-
Qwen4 lineup reported at 2026 Yunqi Conference, but release details remain unknown
A report and event photo identify four names in an upcoming Qwen4 family, while leaving the release...
- You’re up to date All news
04 / Collection
Leaks
All leaks-
OpenAI Codex Pro Max Support Reported in Backend Changes
A report points to ProMax support across Codex server-side systems, but the changes do not confirm...
-
ChatGPT Pro Max: What the Report Shows
A post points to possible higher-tier ChatGPT Pro code references, but the screenshot does not confirm a...
-
Kling 4.0 leaked features: What the report says
An unverified X report describes possible Kling 4.0 video and image features, including 1080p output,...
-
Nano Banana 2.5 Flash Appears in Google Flow
A screenshot shared on X shows Nano Banana 2.5 Flash listed in a Google Flow model selector, but the...
-
Gemini 4 Early Post-Training Release: What’s Reported
A report describes Google’s intention to release an early post-training version of Gemini 4, but the...
-
Claude Sonnet 5.5: What the Report Actually Says
A single X post reports that Anthropic is stealth testing Claude Sonnet 5.5, but the supplied evidence...
-
What the Astra Minor Screenshot Actually Shows
A supplied screenshot appears to list Astra Minor as a standard mainline model with standard safeguards...
-
Claude Opus 5.5 Reported in Claude Code 2.1.280
An X post points to a screenshot listing claude-opus-5-5 in an apparent Claude Code 2.1.280 update, but...
-
GPT-6-Sol Ultrafast Mode: What the Post Says
An X post speculates about a possible GPT-6-Sol Ultrafast mode and a speed of 750 tokens per second, but...
-
GPT-6 Sol, Luna and Astra: What the Azure Report Shows
An unverified X post points to Azure configuration entries for GPT-6 Sol, GPT-6 Luna and GPT-6 Astra...
-
What Is OpenAI’s Rumored Aeon AI Assistant?
A post links the unconfirmed name Aeon to a possible OpenAI assistant, while an excerpt attributed to...
-
What an X Post Says About GPT-6 Sol, Luna and Anthropic 5.5
A single, truncated X-post excerpt reports rumored OpenAI and Anthropic model updates, but does not...
- You’re up to date All leaks
05 / Collection
AI releases
All releases-
Google introduces Gemini 3.8 Flash TTS models for expressive audio
Google has announced Gemini 3.8 Flash TTS and Flash-Lite TTS, with access through the Gemini API and...
-
GPT-6 Sol and Luna: Pricing, Access, and Reported Gains
OpenAI says GPT-6 Sol and GPT-6 Luna bring lower API prices, improved caching, and reported performance...
-
Claude Opus 5.5: What the Announcement Says
Anthropic presents Claude Opus 5.5 as the first model in a new Claude 5.5 family, with reported Fable...
-
Tencent Hunyuan Launches Hy Image3.5 Preview
Tencent Hunyuan has announced a Hy Image3.5 preview with text-to-image and image-to-image generation, up...
-
Xiaomi MiMo-V2.6 Pro and Flash: Reported Benchmarks
Xiaomi announces MiMo-V2.6 Pro and Flash with open model weights and reports mixed benchmark results...
-
Grok 4.7: Pricing, Benchmarks, and Availability
xAI presents Grok 4.7 as a model for longer coding and knowledge-work tasks. Here are its reported...
-
Kimi Code Desktop brings multiple coding agents into one workspace
Kimi describes Kimi Code Desktop as a workspace for managing multiple coding agents, running tasks in...
-
Qwen-Image-2.1: Open-Weight Image Generation and Editing
A grounded overview of Qwen-Image-2.1’s documented image-generation, editing, transparency,...
-
Ternary Bonsai 2 27B Reports 98.2% of Qwen3.8’s Score at 5.9 GB
PrismML says Ternary Bonsai 2 27B is a 5.9 GB model based on Qwen3.8 27B, with 98.2% of its reported...
-
QuiverAI Introduces Arrow 2.0 for Editable SVG Graphics
QuiverAI introduces Arrow 2.0 for generating, editing, vectorizing and animating editable SVG graphics,...
-
Union Alpha is free for a week, says OpenCode
OpenCode describes Union Alpha as a stealth model for agentic coding with image support and says it will...
-
Kimi K2.8 Preview: What Changes in Kimi Code
Kimi K2.8 Preview is fully rolled out in Kimi Code with the existing model ID, adjustable thinking...
- You’re up to date All releases