A user is considering upgrading their GPU setup from a single 3060 to two 5060s to better handle running multiple AI agents. They have found that the Ornith1.5 9B model performs surprisingly well for code-related sub-agent tasks, with their current 3060 capable of running two instances simultaneously at over 25 tokens per second each. The proposed upgrade to two 5060s would allow for four agents to run concurrently with significantly increased speed, provided they can secure the necessary funds. AI
RANK_REASON User discussion about hardware for running AI models, not a release or significant industry event.
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →