AI Coding Rate Limits are RIDICULOUS Now - Here's How You Keep Scaling Anyway

I'm continuing to build out my AI Software Factory and scale up my coding agents, but I'm hitting my rate limits with Claude Code and Codex more and more. This week I hit my limits on the $200 Claude AND Codex plans with days to go before the reset! Very frustrating. The New Opus 5.5 helps but not nearly enough. And now it's obvious that I can't just use the most powerful model for everything. I've been experimenting with mixing models and providers all year, but now it's becoming critical for all of us. So in this video I show you where you actually need the best model and where a smaller, cheaper one gets you basically the same result. I built the same applications three different ways with my factory, and my favorite combo had GPT-6 Astra planning and reviewing while GLM 5.3 Flash wrote every line of code. It cost about 4x less and the apps came out just as good! ~~~~~~~~~~~~~~~~~~~~~~~~~~ - Check out Scrimba Explain - ask a question and get a narrated video lesson back, built while you watch: https://scrimba.com/explain?via=colemedin-explain - Check out Neon's AI Gateway - at-cost access to pretty much every open model, including GLM 5.3 Flash: https://get.neon.com/OU2cOGo ~~~~~~~~~~~~~~~~~~~~~~~~~~ - Join Dynamous for the new Agentic Coding Course 2.0 - the core of the course is now COMPLETE: https://dynamous.ai - The AI Software Factory I used for all the testing (the install prompt is at the top of the README): https://github.com/coleam00/ai-software-factory - Archon, my open source harness builder that runs every workflow in this video: https://github.com/coleam00/Archon - LiveBench, the leaderboard I show in the video: https://livebench.ai - Pi, the coding agent harness I use with Neon's AI Gateway: https://pi.dev - Omnigent, another harness that makes it easy to mix models and providers: https://github.com/omnigent-ai/omnigent ~~~~~~~~~~~~~~~~~~~~~~~~~~ 0:00 My Rate Limits Are Ridiculous Now 1:51 Testing Models With My AI Software Factory 2:50 The Best LLMs for Coding RIght Now 3:34 Big Model Plans, Small Model Builds 4:44 My Favorite Combination Right Now 5:04 Sponsor: Scrimba 6:35 How to Combine Models and Providers 7:27 The Workflow Behind Every Test 8:16 Open Models Through Neon's AI Gateway 10:02 Build 1: Open Models Only 11:03 Build 2: Claude Fable 5.1 Only 12:00 Build 3: GPT-6 Astra + GLM 5.3 Flash 12:52 You Don't Need the Best Model Everywhere ~~~~~~~~~~~~~~~~~~~~~~~~~~ Join me as I push the limits of what is possible with AI. I'll be uploading videos weekly - at least every Wednesday at 7:00 PM CDT!