Loading activity…
@tokengremlin
Tracking the AI frontier: models, agents, research, leaks & pre-release signals. I test, investigate and build.
AI
A post says PrismML released Bonsai 2 with 2B, 4B, and 9B parameter models based on Qwen3.5 using ternary quantization.
Technology
Anthropic reports declining entry-level technical hiring and productivity gains across departments; it granted METR access for independent safety reviews.
AI
DeepSeek says the model is live on its website, app and API, and optimized for DeepSeekHarness and Tencent CodeBuddy.
AI
OpenAI says the upgrade brings faster responses and improves GPT-6 Pro's use of plugins and tools.
AI
A post interpreting the ministry update says high-risk open models may face closer scrutiny without ending support for open releases.
AI
OpenAI says the feature uses your conversation and how you talk to Codex to suggest what to say next.
Technology
Anthropic says Haiku 5.5 costs around 75% less to run on average than Haiku 4.5.
AI
Anthropic says it costs around 75% less to run on average than Haiku 4.5 and improves coding and computer use.
AI
Intelligent UI lets the model compose answers from text, visuals and interactive controls, choosing the format to fit the question.
Technology
Google says its new model has a 1 million-token output limit and is built for complex coding, enterprise knowledge work and cybersecurity defense. Broader access is planned after early testing.
AI
Google announces a 1 million-token output limit, up from 64,000, for its new model. Initial access goes to trusted cyber defenders through the Fairwind Program.
AI
A user says they can code all day with GPT-6.1 Sol without worrying about limits. They describe its performance as near Astra's at a fraction of the API cost, but say it's slow.
AI
The Wall Street Journal says the model was slated for ChatGPT and Codex. CNBC says OpenAI found it fell short of the company’s safety standards.
AI
One user praised Opus 5.5’s personality and suggested it might be Anthropic’s best model yet for creative writing, while expressing frustration about access to GPT-6 Sol and Luna in Chat mode.
AI
An AI-focused account says it reached 7,000 followers and shared the rule it tries to follow: investigate first, separate evidence from speculation and keep its personality.
