What dev communities were talking about this week 👂 The FriendliAI team brings you this week's pulse: 1/ Open-weight models are becoming a real alternative. Deepseek v4 flash scores close to frontier models at ~50x lower cost. even closed-model users are switching 2/ "Open…

Every frontier model speaks its own "language" for tool calling. Instead of hand-building a parser per model, we built a system that auto-generates the spec and runs it through one unified parser. Onboarded GLM-5.2 in days, not weeks. Full breakdown here👇 Article Why We…

If you're building agents, Catch our CBO Brian Yoo speaking at @Ai4Conferences. The Frontier AI Inference Cloud for Agents Aug 5, 11:20 AM, Delfino Ballroom 4104

Your response_format validates. The output is still useless. 🤷 Constrained decoding only masks invalid tokens — it doesn't pick the right one. Structure ≠ intent. Bonus: up to 21% faster on FriendliAI via speculative decoding. Full breakdown 👇 Article Structured Output…

Everything you want your inference stack to be. Inference is Friendli → friendli.ai Choose your Model → friendli.ai/models