Loading activity…
@nvidiaai
Teaching your AI new tricks.
AI
A Nvidia GTC attendee in Berlin said they started using GPUs for deep learning as a CIFAR professor in 2009.
AI
ModelScope says its ms-swift toolkit supports inference and fine-tuning for the 120B-parameter multimodal model.
AI
The launch post claims the diffusion LLM outperforms GPT-6 Luna and Haiku 5.5 on quality.
AI
NVIDIA AI shares a broadcast link it frames around European AI builders.
AI
Adaption AI says its co-founder will discuss its prediction that learning from users' work, not scale, will define next-generation systems.
AI
NVIDIA AI says its five-minute video walks through Dynamo deployment with scalable workers and an OpenAI-compatible endpoint.
AI
Scale says its post-trained model matched Skydance’s frontier-model accuracy at 40x lower inference cost.
AI
River says the model beat Kimi K2.6 on its text-to-SQL evaluation before training.
AI
Bespoke Labs says the evaluation covered 22 CPU tasks, each with a four-to-six-hour limit and one graded submission.
AI
Underdog says its Saluki 27B model fits in under 8 GB and performs on par with Claude Opus 4.6.
AI
Reflection says the release includes Mirror CLI, its open-source training harness, and collaborations with 40-plus partners.
Technology
NVIDIA says its new Nemotron 3.5 Bolt 24GB fits into 24GB of memory.
AI
Michael Dell says the models are available with Hugging Face, framing the additions as more choice for customers.
AI
Perplexity's README says SPACE uses isolated Firecracker microVMs and an HTTP API to run commands and collect output.
AI
Nvidia says its open model combines reasoning and tool use with text, image, document and video understanding.
AI
AfterQuery says post-training Nvidia's model with Hebbia raised 'perfect solved' from 47% to 74%.
AI
finkd says the model runs well on new Mac Studios, crediting work with Apple and John Ternus.
AI
ArtificialAnlys says that’s 1.8 times Super VL’s 46,000, but 28% fewer than Nemotron 3 Super’s 114,000.
AI
PrimeIntellect says it built 2,500 hard enterprise tasks with NVIDIA, modeled on AutomationBench.
AI
BFL says it runs on a single consumer GPU and offers native 4K generation, bounding boxes and pixel-perfect editing.
