Open-source models have reached a new level of intelligence; our own engineers use models like GLM-5.2 for tasks like kernel optimization. Thanks @latentspacepod for having @waterloo_intern and @philipkiely on to talk about how we optimize models for the highest throughput…

One of our engineers asked @poolsideai's Laguna S 2.1 to transform a 715-file C++ game from neon cyberpunk into an Ancient Greek aesthetic. It orchestrated 3 different models, refactored the code, and produced a playable build. baseten.co/blog/laguna-s-21-…

Yes, we will have Qwen3.8-Max.

@opencode models.dev PR is merged, so now you can use it in opencode after adding your @baseten api key. or in claude/codex with b10-switch!

Try DeepSeek V4 Flash on our Model APIs today. - 80–98% cheaper than other frontier models - Comparable intelligence - 1M token context window Use it here: baseten.co/library/deepseek-…