Qoder Launches Community Boost With Up to 4,096 SOTA Credits for Sonus
Qoder launched a Community Boost campaign offering 1,024 SOTA Credits for Sonus, with tiered daily boosts unlocking up to 4,096 credits. Pro users join automati…
Global AI Digest
-
-
NVIDIA and Hugging Face Launch SimReady Central Library with 2,000+ Simulation-Ready 3D Assets
NVIDIA announced the SimReady Central Library on Hugging Face, offering over 2,000 open-source simulation-ready 3D assets for robotics simulation and AI factory…
-
GMI Cloud Offers 40% Off DeepSeek V4.1 Flash 1017 for 10 Days
GMI Cloud is offering 40% off DeepSeek V4.1 Flash 1017 through its API for 10 days with unlimited usage.
-
Japan's PM Urges Nationwide PHP Patch for Two Critical Vulnerabilities
Japan's Prime Minister and national cybersecurity office asked website operators to apply a PHP security patch released at 11 PM JST on October 21, warning that…
-
Railway Breach Highlights the Black Market for Stolen AI API Keys
dax notes that the Railway breach appears to be another case of stolen Anthropic and OpenAI API keys, arguing tokens function like cash because distilleries wit…
-
GMI Cloud CLI Generates 5-Second Video in ~27 Seconds With Upfront Pricing
A recorded GMI Cloud CLI session shows /image and /video commands with prices shown before execution: a $0.035 image and a 5-second video at $0.022 per second,…
-
Recursive_SI's Flint 40B Coding Model Lands on OpenRouter, Free at Launch
Recursive_SI's first model, Flint, is an open-weight 40B MoE (~2.5B active) built for coding agents with 262K context, reasoning and tool calling, now available…
-
Recursive SI Releases Flint, a Fast Local Coding Model
Recursive SI released Flint, its first model, a fast local coding model built with its automated research system Eureka Machine, claiming up to 2x faster infere…
-
MiniMax M3.1 Flash Gains Inferact as API Launch Partner
Inferact joins as an API launch partner for MiniMax M3.1 Flash, bringing vLLM serving expertise for agent workloads; the model scores 29.6 on Agents' Last Exam…
-
MiniMax H3 Runs on TPUs via SGLang-TorchTPU
MiniMax acknowledges H3 being brought to TPUs through SGLang-TorchTPU, a TPU plugin for SGLang built with Google and Meta PyTorch teams, enabling efficient open…
