You can deploy this model and start earning money today!

amanwalksdownthestreet/
Llama-3.3-70B-Instruct-MXFP8

Name: Llama-3.3-70B-Instruct-MXFP8
Availability: InStock

MXFP8 quantization of meta-llama/Llama-3.3-70B-Instruct

Quantized with llm-compressor 0.13.1.dev2+g20499a97b.d20260813 (MXFP8 W8A8, lm_head and embeddings retained at original precision).

Requires Blackwell (SM100+). Older GPUs such as Ada, Hopper, etc. cannot run this model — MXFP8 matmul needs Blackwell's native microscaling support.

Serving with vLLM

bash

vllm serve amanwalksdownthestreet/Llama-3.3-70B-Instruct-MXFP8 \
    --max-model-len 41184

Weights occupy ~70 GiB. On a single 96 GB Blackwell card, KV cache headroom supports a context length of up to 41184 tokens (the model's native 131072 requires ~40 GiB of KV cache — multi-GPU territory).