Hugging Face Now Routes Open-Weight Models Through OpenRouter
Hugging Face has switched its language model inference to OpenRouter, so users can call served open-weight models on the Hub with one Hugging Face token and automatic provider…
More freedom.
More control.
New AI releases, creative tools and what you can actually download, change and use.
118 stories
Hugging Face has switched its language model inference to OpenRouter, so users can call served open-weight models on the Hub with one Hugging Face token and automatic provider…
An arXiv preprint dated November 10 2026 introduces SteerDistill, a closed-loop distillation method using a full-precision teacher to correct quantized LLM agent failures…
A November 10 2026 arXiv preprint introduces SLIC, a reinforcement-fine-tuned 4-billion-parameter model that runs locally and optimizes analog circuits in closed-loop simulator…
An arXiv preprint from November 4 2026 introduces TileQuant, a training-free method that assigns 16-bit or 8-bit precision to small 2D tiles inside LLM weights. It delivers up…
Ai2 receives multimillion-dollar funding from Google.org to develop NeuralGCM+, extending the hybrid AI-physics model with coupled ocean and sea-ice components for reliable…
Collinear AI made CWE-bench v2 public with 100 reproducible tasks that test whether AI models can find and patch vulnerabilities in actual open-source code without breaking…
Nextcloud has awarded OLMo 3 its top green Ethical AI rating, citing the model's fully open training data, weights, and reproducible build process. Organizations can now run…
Google released Gemini 4 Argon globally for coding, analysis, and robotics while announcing Gemini 4 Carbon in limited pre-release testing with government and cyber defenders.…
A new arXiv preprint presents TICL³, which combines a shared linear model with local corrections guided by a tabular foundation model. This lets the same coefficients both…
An arXiv preprint released October 25, 2026 tests Llama 3 8B and Qwen 2.5 7B zero-shot and after QLoRA fine-tuning against TF-IDF and CodeBERT baselines on 20,000…
A preprint dated October 22 2026 introduces Sensorimotor Alignment (S²), a training-free offline way to measure how well a model's sensory representations align with physical…
Enterprises can now run Mistral models and train custom versions inside Cloudera’s hybrid platform, keeping data, training, and inference fully under their control across…