Frontier open-source MoE model built for OpenClaw agents
Step 3.5 Flash is StepFun’s 196B sparse MoE model that activates only 11B parameters per token. It delivers frontier reasoning and strong agentic performance with high efficiency. Seamless native OpenClaw integration makes it one of the best open models for running serious agents right now.
Step 3.5 Flash has been out for a few weeks and has quickly become one of the strongest open models for real agentic workflows. 196B sparse MoE with just 11B active per token, MTP-3 giving up to 350 tok/s on coding, solid 74.4% SWE-bench, and clean long-context handling.
Right now you can just plug it into your Claw using the free quota on OpenRouter or the official StepFun API. Super easy to test and it performs really consistently in long agent loops.
Zac Zuo
Hi everyone!
Step 3.5 Flash has been out for a few weeks and has quickly become one of the strongest open models for real agentic workflows. 196B sparse MoE with just 11B active per token, MTP-3 giving up to 350 tok/s on coding, solid 74.4% SWE-bench, and clean long-context handling.
The OpenClaw support is seamless. Over the last couple days Step 3.5 Flash has even been #1 in daily OpenClaw token consumption on @OpenRouter.
Right now you can just plug it into your Claw using the free quota on OpenRouter or the official StepFun API. Super easy to test and it performs really consistently in long agent loops.