Local model performance on your hardware

generation:

25.0 tokens/s

prompt processing (prefill):

3557.0 tokens/s

quant:

NVFP4 (NVFP4)

kv:

8-bit

User benchmarks Qwen3.8-27B at 24.96 t/s generation and 3557 t/s prompt processing on a Jetson AGX Thor 128GB, using the Mjolnir vLLM image with the FA4 GEMV decode kernel. Setup is vLLM 0.30.0 with NVFP4 weights and an 8-bit KV cache at 8K context, single request. The GEMV kernel is default-on and dispatches only on M=1, head_dim=256, GQA shapes. The GEMV leg wins at c=1 but regresses at c=4 (57.74 t/s at 8K context, −15.4% versus stock vLLM), which the user attributes to an open investigation. Stock vLLM reaches 24.43 t/s and 2468 t/s prompt at the same setting.

generation:

139.1 tokens/s

quant:

NVFP4 (NVFP4)

User reports Qwen3.6-35B-A3B-NVFP4 at 139.1 tok/s on a single NVIDIA Jetson AGX Thor with 117 GB unified memory. Setup is vLLM built from source for sm_110a with DFlash speculative decoding (12 tokens), marlin MoE backend, flash_attn attention backend, 65536 context length, and 0.78 GPU memory utilization. The post also benchmarks Qwen3.5-4B-NVFP4 at 155.8 tok/s, Qwen3.6-27B-NVFP4 at 50.1 tok/s, and Qwen3.5-122B-A10B-NVFP4 at 52.6 tok/s, all at concurrency 1. The 122B requires cutlass MoE and TRITON_ATTN due to a Marlin crash at 256 experts, and achieves 27-42 tok/s with DFlash versus 10.9 tok/s autoregressive.