FriendlyElec NanoPi R6S (Rockchip RK3588S, 8GB)
Compare FriendlyElec NanoPi R6S (Rockchip RK3588S, 8GB) against another machine in the Arena
Reference: Rockchip on Wikipedia
- Class
- SoC
- Memory
- 8 GB LPDDR4X
- Runs (Q4_K_M)
- —
- Bandwidth
- 34 GB/s
- Released
- 2022-10-28
- Price (US)
- $212 new as of 2026-08
- Price (Canada)
- CA$229.99 new sourced, as of 2026-08
Sourcing and disambiguation notes
Not a mainstream-llama.cpp NPU: the RK3588S 6 TOPS engine needs the separate RKLLM runtime, so the one measured data point below is CPU-only, and it is a 2023 LLaMA-1 7B run -- treat it as a lower bound rather than as representative of current builds on this chip. The board has risen to about $212 direct from FriendlyElec, from $119 at its 2022 launch. Both the US and Canadian prices are the same overseas seller shipping direct from China rather than domestic retail, so the unusually small gap between them is a direct-import price on both sides, not a Canadian discount.
Specification sources
- https://liliputing.com/nanopi-r6s-is-a-single-board-pc-with-rk3588s-8gb-ram-and-three-ethernet-ports-for-119/ — press, 2022-10-28
Measurements
Decode is token generation — the speed you feel while an answer streams. Prefill is prompt processing — the wait before it starts. Why bandwidth predicts decode speed.
No records yet for this hardware. Know of a published benchmark on it? Submit the link.
What verified, single-source and estimated mean, and the same rows with every filter and sort in the benchmarks explorer.
Speed over time
What it can run
Fit is arithmetic, not a measurement — how it is computed. Against 8 GB; models that fit are listed largest first.
No modeled quant fits in 8 GB.