Global AI Digest

2,700 posts · Updated 4 hours ago

Tracker Codex account reset Check whether Codex has reset today, and when it last did. Directory Free LLM APIs Every free AI model API tier, compared in one place. Briefing AI daily briefing The day's most important AI news, condensed into ten items.

OpenCode Browse Outperforms Vercel's agent-browser in Benchmark

A user benchmark shows OpenCode Browse running 40% faster than Vercel's agent-browser with 81% fewer input tokens and 33% fewer steps, using GPT 6 Astra and no…

OpenCode 4 hours ago #OpenCode#agent-browser#benchmark
GB300 and B300 Rack Pricing and Availability Shared

A user shares pricing for GB300 and B300 racks: GB300 at $9-10/hr with under one-year commitment and one-week readiness, B300 at $6-7/hr on the same terms.

dax 5 hours ago #GB300#B300#GPU pricing
AI-Assisted Porting to Rune Framework Shows Strong Early Results

A developer reports that AI successfully ported a TUI app to the Rune framework with good results, surprising the project's creator given it launched only a wee…

dax 5 hours ago #AI#code migration#Rune
CommandCodeAI's North Star Metric: Cache Hit Rate at 100T Token Scale

Ahmad Awais says CommandCodeAI's internal RSI engine uses cache as its north star metric, enabling $100 worth of inference for $10 and surfacing 2-3M code impro…

Ahmad Awais 9 hours ago #AI#inference#cache
OpenCode Browse Can Run Agents and Browsers on Different Machines

OpenCode Browse now supports running the agent and the controlled browsers on separate machines, so a remote server can orchestrate browsers on a desktop and la…

dax 9 hours ago #OpenCode#AI agent#browser automation
LiteLLM Weekly Update: Routing Fully Migrated to Rust, New Model Aliases, and Skills Hub

LiteLLM's weekly update covers a full migration of routing to Rust, new model support, a Codex model picker, -latest model aliases, and the Skills Hub.

LiteLLM (YC W23) 13 hours ago #LiteLLM#Rust#LLM routing
DeepSeek V4.1 Flash Hits 1000 TPS in Command Code

Command Code has launched DeepSeek V4.1 Flash 1017 Ultrafast with a 1M token context window and over 1000 tokens per second throughput, alongside GLM 5.3 Flash…

Ahmad Awais 13 hours ago #DeepSeek#Command Code#LLM inference
LongCat-2.5 Now Free for One Week on Nous Research Portal

Meituan LongCat announced that its LongCat-2.5 multimodal model is available for free for one week on the Nous Research Portal, where users can test it together…

Meituan LongCat 14 hours ago #LongCat-2.5#Nous Research#Hermes Agent
OpenCode Browse Lets Agents Control Your Real Browser Locally

OpenCode Browse is now available, allowing an agent to control the user's real browser with existing tabs, logins, and extensions, while keeping data on-device…

dax 14 hours ago #OpenCode#AI agent#browser automation
OpenCode Browse Lets Agents Control Your Real Browser

OpenCode released OpenCode Browse, a feature that lets an agent control the user's real browser, including existing tabs, logins, and extensions.

OpenCode 14 hours ago #AI agent#browser automation#OpenCode