LiteLLM Adds Per-Customer Usage and Spend Tracking for AI Agents
LiteLLM Customer Management tracks usage and spend per customer across organizations, teams, and members, and supports pricing models such as premium model tier…
LiteLLM Customer Management tracks usage and spend per customer across organizations, teams, and members, and supports pricing models such as premium model tier…
Google is rolling out computer use in the Gemini app for Mac, letting Gemini operate across folders, apps, and the web in a background window while the user kee…
Perplexity released pplx-decider-v1.3 and pplx-decider-flash-v1.3, adding a 'locate' feature that extracts spans by type or instruction, available via its Decis…
A developer integrated Gemini 4 models into Amp through GMI Cloud's API, mapping the models so Amp's existing prompt optimizations apply directly, and noted Car…
Perplexity is keeping its Decider-v1.3 and Flash decision models free on OpenRouter for another three weeks, citing full-stack ownership and a ~20% share of req…
Command Code introduced Codemode, a feature that lets agents chain MCP tool calls into a single script. In a demo, model round trips fell from 44 to 3 and input…
OVHcloud released Olala mini 7A1B, a 7-billion-parameter open-source model built on a new architecture, trained on European public supercomputers, and able to r…
Airbnb Engineering benchmarked Gemma 4 E2B for policy classification, reaching 0.902 max F1 with 35ms median latency on a single NVIDIA RTX Pro 6000 GPU at abou…
Nous Research announced a Hermes adapter for Alibaba's Qwen3.8-27B that improves SWE-Bench Pro and Terminal-Bench scores while cutting token usage by 3.5-4.4x w…
Ollama cofounder and CEO shares that conversations with CIOs and engineering leaders at enterprises have shifted from asking where to run models to asking where…