OpenClaw is seeking early users to test their open-source model inference plans, sold by concurrency slot with high throughput and no shared pool, in exchange for free access and feedback.
Hey everyone 👋 We're working on **OpenClaw model plans**. Open-source models sold by concurrency slot, not by token. *Nothing new, right? Well...* each slot is yours: from 100 tokens/s and a 500M+ daily token cap depending on model, no shared pool, no speed degradation under 24/7 load. Built for the exact problem that keeps showing up in this sub. The current stack we're validating: * Qwen 3.6 35B-A3B * Qwen 3.6 27B * Gemma 4 31B * Qwen 3.5 122B-A10B * DeepSeek v4 Flash This is our early-access lineup. Which models stay, which get cut, which we add. That'll come from real demand. Looking for a handful of OpenClaw users running real agent loops who'd be up for: * Virtual coffee, 15–30 min with me. Relaxed, informal, no slides * Free access to the tier that matches your workload, for a month * Trying it on something real and telling us honestly what sucks. Website, pricing, docs, the inference itself, anything If our thing is worse than what you're already using, I want to know why. If you're running OpenClaw at any serious volume and this sounds interesting, drop a comment or DM. I'll respond to all of them, and we'll set up the coffee time. Thanks 🙏
A user asks the community about their real-world experiences with OpenClaw, seeking honest feedback on common workflows, cool automations, frustrations, and setup configurations.
The blog post describes using local open-weight models like Gemma and Qwen in an agent harness to automatically triage issues and pull requests in the OpenClaw repository, enabling real-time notifications without relying on costly closed API models.