Can a Rockchip RK3576 run Qwen3.5 0.8B?

Yes, at a comfortable chat speed. Measured generation speed: 13.18–18.79 tok/s across 3 public runs.

Rule of thumb: 10+ tok/s feels like a chat; 4–10 is usable for assistants and automations; under 4 is batch-job territory. Speeds depend on quantization, runtime, context length and cooling. Check each source.

BoardSizeQuantRuntimeGen tok/sPrompt tok/sTTFT msSource
Rockchip RK3576
Rockchip RK3576 (vendor reference board)
800Mw4a16RKLLM v1.3.118.79–1369.31source ↗
commit 2026-09-29
Rockchip RK3576
Rockchip RK3576 (vendor reference board)
800Mw4a16_g128RKLLM v1.3.117.88–1392.09source ↗
commit 2026-09-29
Rockchip RK3576
Rockchip RK3576 (vendor reference board)
800Mw8a8RKLLM v1.3.113.18–1342.54source ↗
commit 2026-09-29

Qwen3.5 0.8B on other boards

Buy a Rockchip RK3576

Ran Qwen3.5 0.8B on a Rockchip RK3576? Submit your numbers →

Stay in the loop.

New boards, new runtimes, new numbers. One email a month, no hype.