Nemotron-Labs-Sigma, 3B and 8B continuous diffusion language models with 16-D embeddings and 3.46x speculative decoding ...
Baseten's bison-1 decision model, built on Gemma 4, ranks first on Artificial Analysis Decision Index v2.1 with 76.
Cmpnd released dspy-wsp, a prototype DSPy optimizer for weight-space programming. A frontier LLM reads a small model’s ...
FlashMap is a speculative decoding drafter from Polymathic AI, Princeton and USC ISI researchers. It drafts 16 tokens jointly ...
NVIDIA released Nemotron 3.5 Transcribe, an open-weights speech LLM with 4.3B active parameters. It transcribes 40 languages, ...
Meta FAIR’s Self-Improving Memory lets Kimi-K2.6 write its own LOAD and OFFLOAD memory prompts, then distills them into its weights. Optimized prompts score 56.0 versus 44.6 for the best hand-written ...
Meta FAIR’s Self-Improving Memory lets Kimi-K2.6 discover its own memory prompts, scoring 56.0 versus 44.6 for hand-written ...
Perplexity Orchestrator API unifies OpenAI, Anthropic, open-weight models; routing lifted Fable 5.2 from 38 to 41 ...
AI attackers are speeding up: the 4-pillar incident response playbook for AI platforms, plus Fireworks AI and Klue.
RadixArk released DaisyRose-27B, an Apache 2.0 agentic coding model post-trained from Qwen3.8-27B with reinforcement learning only. It more than doubles Terminal Bench 4.1 pass@1 to 12.0 and runs on a ...
Google Deeepmind's Gemini 4 Neon offers 1M context at $3/$15 per million tokens and tops AA’s Coding Agent Index.
TinyFish published two engineering research articles on agentic search. One defines a 7-phase production architecture with ...