FULL ARCHIVE
Follow the signal.
Browse every selected story by topic, beyond the twenty-item live edition.
MODELS
56 STORIESMODELS
Scaling Indic Pretraining: 2.8 Trillion Tokens and an Open Data Generation Pipeline
NVIDIA details 2.8 trillion Indic-language pretraining tokens and releases an open data generation pipeline.
MODELS
Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint
IBM announced Granite Switch 4.1 as a single endpoint for chat, decisions, and speech.
MODELS
GPT-Live-1.1 has new voices and better accents.
OpenAI posted a new video announcing GPT-Live-1.1 voices and improved accents.
MODELS
Apertus 1.5 Release Notes
Hugging Face posted the official release notes for Apertus 1.5 today.
MODELS
TypeSafe AI's Jev 1.14 is now available on AI Gateway
Vercel added Jev 1.14 to AI Gateway with image input and long-context mode.
MODELS
Integrating Netflix’s Foundation Model into Personalization applications
Netflix details practical integration patterns for its personalization foundation model in production systems.
MODELS
The Lifecycle of LLM-as-a-Judge: Building, Aligning, and Monitoring at scale
Netflix published its current framework for building and monitoring LLM judges at scale.
MODELS
Nemotron 3.5 Lightning now available on AI Gateway for free
Vercel added NVIDIA’s Nemotron 3.5 Lightning Free to AI Gateway today.
MODELS
We tested Gemini 4 Argon for 1,000 hours
Google Cloud published a new video describing 1,000 hours of Gemini 4 Argon testing.
MODELS
Run fast, private agents on-device with LFM2.5-24B-A2B
Hugging Face published Liquid AI’s LFM2.5-24B-A2B launch for fast private on-device agents.
MODELS
Gemini 4 Argon for Robotics
DeepMind announced Gemini 4 Argon for robotics in a dedicated launch post.
MODELS
Gemini 4 Argon: Built to solve your hardest problems
DeepMind introduced Gemini 4 Argon as a new model for hard problems.
MODELS
Judge a Transaction by the Company It Keeps — Somya Rai & MD Sharique, EXL
EXL discusses transaction analysis using relational context instead of single-event features.
MODELS
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space analyzes MSL’s Muse/Wonder model as a newly credible frontier lab effort.
MODELS
Introducing Clef-omni with full multimodality, plus a faster Clef and a cheaper Clef-flash
Cloudflare launched Clef-omni and announced faster Clef inference with lower Clef-flash pricing.
MODELS
Liquid AI's d1 is available on AI Gateway
Vercel added Liquid AI’s d1 decision model to AI Gateway with vision support.
MODELS
Step 5 Preview now available on AI Gateway
Vercel added StepFun’s Step 5 Preview model to AI Gateway with a 1M-token context window.
MODELS
Introducing Claude Haiku 5.5
Anthropic just launched Claude Haiku 5.5 with new pricing and tokenization tradeoffs.
MODELS
Introducing Mistral Large 4: Le chonk
Mistral previewed Large 4 via API and says open weights are coming later this month.
MODELS
EmbeddingGemma 2: an open, lightweight multimodal embedding model
DeepMind announced EmbeddingGemma 2 as an open lightweight multimodal embedding model.