How much VRAM does Qwen3.8 2.4T A95B need?
Qwen3.8 2.4T A95B needs 1,381 GB for its weights at 4-bit, 1,384 GB with an 8K context and 1,386 GB with 32K. At 8-bit, 2,423 GB.
- Parameters
- 2.4T
- Active per token
- 95B
- Longest context
- 256K
- Released
- Aug 2026
- License
- Qwen3.8-Max License
- Strongest at
- Chat, Coding, Agents
Memory at each precision
| Precision | Weights | With 8K | With 32K | With 128K |
|---|---|---|---|---|
| 16-bit | 4,556 GB | 4,559 GB | 4,561 GB | 4,570 GB |
| 8-bit | 2,421 GB | 2,423 GB | 2,425 GB | 2,434 GB |
| 6-bit | 1,868 GB | 1,871 GB | 1,873 GB | 1,881 GB |
| 5-bit | 1,620 GB | 1,623 GB | 1,625 GB | 1,634 GB |
| 4-bit | 1,381 GB | 1,384 GB | 1,386 GB | 1,394 GB |
| 3-bit | 1,113 GB | 1,116 GB | 1,118 GB | 1,127 GB |
| 2-bit | 854 GB | 857 GB | 859 GB | 868 GB |
| 1.6-bit | 470 GB | 472 GB | 474 GB | 483 GB |
One conversation, 16-bit context cache, runtime overhead included.
What runs it
| Machine | Best fit | Context | Speed | First reply |
|---|---|---|---|---|
| RTX 3060 12 GB | Too big | |||
| RTX 4060 Ti 16 GB | Too big | |||
| RTX 3090 | Too big | |||
| RTX 4090 | Too big | |||
| RTX 5090 | Too big | |||
| 2× RTX 3090 | Too big | |||
| RTX PRO 6000 Blackwell | Too big | |||
| M4 Max 128 GB | Too big | |||
| M5 Max 128 GB | Too big | |||
| Ryzen AI Max+ 395 128 GB | Too big | |||
| DGX Spark 128 GB | Too big | |||
| M3 Ultra 256 GB | Too big | |||
| M5 Ultra 512 GB | Too big | |||
| H100 SXM | Too big | |||
| DGX Station GB300 | 1.6-bit, spills | 256K | 35 tok/s | 4.5 s |
| 8× H200 NVL | 2-bit | 256K | 61 tok/s | 0.8 s |
Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.
How to run it
- llama.cpp
llama-server -hf unsloth/Qwen3.8-2.4T-A95B-GGUF- vLLM
vllm serve Qwen/Qwen3.8-2.4T-A95B
Questions
How much VRAM does Qwen3.8 2.4T A95B need?
Qwen3.8 2.4T A95B needs 1,381 GB for its weights at 4-bit, 1,384 GB with an 8K context and 1,386 GB with 32K. At 8-bit, 2,423 GB.
Can Qwen3.8 2.4T A95B run on an RTX 4090?
Not on one card. The smallest machine here that runs it is the 8× H200 NVL.