@ekzhang1 published a Jev-compatible API that runs Qwen3.6-35B-A3B with SGLang radix cache so prefill can be reused. They say 64 tasks finished in under 1 second.

Eric Zhang (@ekzhang1) published this on 2026-09-17. @ekzhang1 published a Jev-compatible API that runs Qwen3.6-35B-A3B with SGLang radix cache so prefill can be reused. They say 64 tasks finished in under 1 second. Inspired by @typesafeai , here is a Jev-compatible public API to play with It runs a comparable open model (Qwen3.6-35B-A3B), and just uses SGLang radix cache to preserve the prefill reuse / really fast parallel systemone generation - 64 tasks in <1s. How Jev or TypeSafe is used, in the post's own words: “Inspired by @typesafeai , here is a Jev-compatible public API to play with It runs a comparable open model (Qwen3.6-35B-A3B), and just uses SGLang radix cache to preserve the prefill reuse / really fast parallel systemone generation - 64 tasks in <1s.” The post does not name a separate audience. It is for people reading what @ekzhang1 shows here (it links to github.com; it includes media).