Qwen3.6-35B-A3B-NVFP4 vs Meta-Llama-3.1-70B-Instruct-FP8
Specs, VRAM requirements and download trends — updated 10 November 2026.
nvidia · A · Open weights
Params
18.7 B
Context
—
30d
4.0 M
apache-2.0 · commercial OK
RedHatAI · B · Open weights
Params
70.6 B
Context
131,072
30d
142 K
llama3.1 licence · conditional
Specification comparison
Differences are highlighted; identical values are muted.
| Specification | Qwen3.6-35B-A3B-NVFP4 | Meta-Llama-3.1-70B-Instruct-FP8 |
|---|---|---|
| Parameters | 18.7 B | 70.6 B |
| Architecture | Qwen3_5MoeForConditionalGeneration | LlamaForCausalLM |
| Context length | — | 131,072 |
| Licence | apache-2.0 | llama3.1 |
| Commercial use | Allowed | Conditional |
| Languages | — | 8 |
| Downloads 30d | 3,971,809 | 141,993 |
| Downloads all time | 42.1 M | 2.1 M |
| Quantizations on the Hub | SAFETENSORS | SAFETENSORS |
| Gated | No | No |
| First seen on the Hub | 2026-05-27 | 2024-07-23 |
Download trend
Daily snapshots, last 90 days (13 Aug – 10 Nov)
12.6 M139 K
Qwen3.6-35B-A3B-NVFP4
Meta-Llama-3.1-70B-Instruct-FP8
VRAM side by side
On RTX 4090 · 24 GB · 8K context unless noted
Qwen3.6-35B-A3B-NVFP4 · fp16
29.1 GB / 24 GB
Meta-Llama-3.1-70B-Instruct-FP8 · fp16 @ 128K ctx
249.8 GB / 24 GB