Loading activity…
@nvidiaai
Teaching your AI new tricks.
AI
PrimeIntellect says Prime Inference's 96.9% cache hit rate brings effective input costs to $0.024 per million tokens.
AI
The SETI Institute and Breakthrough Listen use NVIDIA Holoscan to process live data; agents can calibrate the telescope.
XAI
NVIDIA designed the 9.6-million-parameter multilingual model to address noise, echo, clipping and other recording problems.
XAI
Mistral announced the weights after a public preview, but its linked main Hugging Face listing showed “Upcoming release.”
AI
Samaya says Sterling scored 58.9% versus Claude Fable 5.1’s 55.9%, at $0.61 versus $6.89 per question.
AI
NVIDIA describes speaker attribution and translation in one checkpoint; SiliconANGLE reports a demo and downloadable checkpoint.
X
Technology
SiliconANGLE reports the open-weight model supports 40 languages, labels up to eight speakers and answers questions about recordings.
XAI
NVIDIA says new Data Factory blueprints can generate synthetic training data for driving and inspection tasks.
Technology
Nvidia says Reachy Mini uses Bolt for conversation and Cosmos for vision; both run locally on Jetson Orin Nano 8GB.
XAI
Llama.app says the free, open-source app recommends models that fit your computer and installs them with one click.
AI
In NVIDIA's Reachy Mini demo, Bolt handles conversation while Cosmos 3 Edge supplies vision on Jetson Orin Nano 8GB.
XAI
NVIDIA says applicants must have completed their first Ph.D. year when applying; a summer 2027 in-person internship is mandatory.
AI
Reflection AI says Beam supports Axolotl, NeMo RL and SkyRL; recipes are on its website and in its cookbook repo.
AI
NVIDIA says its Embed models search text and images, while Parse extracts structured content from document pages.
AI
An organizer said Joe Nemotron would join the October 21, 2026, stream to discuss local AI.
AI
In an Oct. 21 post, @rronak_ attributes a “frontier-level capability” claim to Jensen Huang.
Technology
Nvidia's design pairs an open-source runtime with Vera computing; BlueField-4 monitoring is an optional safeguard.
YouTubeAI
Mistral says the companies plan open physics-model datasets and benchmarks, plus evaluations of engineering AI agents.
AI
NVIDIA claims DSX POD can cut deployment at a prepared site from six months to 14 days.
AI
Together AI says engineers from n8n, Nvidia and its team spoke at the event, which drew audience questions.
