Can a Radxa ROCK 5B+ (RK3588) run Gemma 3 1B IT?
Gemma 3 1B IT speed by board
- Radxa ROCK 5B+ (RK3588): 17.0 tokens/s best (Q8_0, rk-llama.cpp fork (NPU)), 1 test
- Raspberry Pi 5: 10.4 tokens/s best (Q4_K_M, llama.cpp b9307 (CPU, OpenBLAS)), 1 test
FAQ
Which edge board runs Gemma 3 1B IT fastest?
Of the boards tested, Radxa ROCK 5B+ (RK3588) is fastest for Gemma 3 1B IT: 17.0 tokens/s generation (Q8_0, rk-llama.cpp fork (NPU)). Next is Raspberry Pi 5 at 10.4 tokens/s.
How fast does Gemma 3 1B IT run on Radxa ROCK 5B+ (RK3588)?
Best reported generation speed for Gemma 3 1B IT on Radxa ROCK 5B+ (RK3588) is 17.0 tokens/s (Q8_0, rk-llama.cpp fork (NPU)), from 1 sourced test. Source: Mojo24x7/rk-llama.cpp (https://github.com/Mojo24x7/rk-llama.cpp).
How fast does Gemma 3 1B IT run on Raspberry Pi 5?
Best reported generation speed for Gemma 3 1B IT on Raspberry Pi 5 is 10.4 tokens/s (Q4_K_M, llama.cpp b9307 (CPU, OpenBLAS)), from 1 sourced test. Source: davidscarth (edge-ai-study) (https://github.com/davidscarth/edge-ai-study/blob/main/RASPI5.md).
| Board ↕ | Model ↕ | Runtime & setup ↕ | Gen t/s ↕ | Price ₹ (India) ↕ | Source ↕ |
|---|---|---|---|---|---|
| Raspberry Pi 516 GB RAM | Gemma 3 1B IT1B params · Q4_K_M | llama.cpp b9307 (CPU, OpenBLAS)NVMe (official 512GB SSD kit) · Active Cooler | 10.4prompt 60.2 t/s | ₹34,191 Buy0.31 t/s per ₹1k | Community davidscarth (edge-ai-study)2026-05-27 |
| Radxa ROCK 5B+ (RK3588)16 GB RAM | Gemma 3 1B IT1B params · Q8_0 | rk-llama.cpp fork (NPU) | 17.0prompt 201.6 t/s | No India listing yet | Community Mojo24x7/rk-llama.cpp2026 |
Each row shows only what its source reported: RAM, quant, prompt speed and setup appear under the main value when published. Rows with full details are listed first, fastest first. Source badges: Community, Vendor, Reproduced; ⚠ marks a speed far from similar results (how we check).