Can I run DeepSeek V4.1 Flash locally?
DeepSeek V4.1 Flash by deepseek-ai has 763B parameters (188B active per token, 384 experts, 6 used per token). At 4-bit with an 8K context it needs about 436 GB of GPU memory. It fits on 1 of 65 GPUs and Macs we track.
This model uses a newer compressed or sparse attention design. Our KV-cache numbers use the standard formula, so treat long-context figures as an upper bound.
Memory needed by quant and context length
GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use parameter count × bits per weight. How we calculate.
| Quant | Quality | 4K ctx | 8K ctx | 32K ctx | 128K ctx |
|---|---|---|---|---|---|
| 2-bit (Q2_K) | rough | 268 GB | 268 GB | 270 GB | 278 GB |
| 3-bit (Q3_K_M) | usable | 348 GB | 349 GB | 351 GB | 359 GB |
| 4-bit (Q4_K_M) | good, most popular | 435 GB | 436 GB | 438 GB | 446 GB |
| 5-bit (Q5_K_M) | very good | 507 GB | 508 GB | 510 GB | 518 GB |
| 6-bit (Q6_K) | near lossless | 584 GB | 584 GB | 586 GB | 595 GB |
| 8-bit (Q8_0) | effectively lossless | 756 GB | 757 GB | 759 GB | 767 GB |
| 16-bit (F16/BF16) | original weights | 1423 GB | 1423 GB | 1425 GB | 1433 GB |
Which GPUs and Macs run DeepSeek V4.1 Flash?
Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).
Prefer an API?
DeepSeek: DeepSeek V4.1 Flash is on OpenRouter at $0.14 per 1M input tokens and $0.42 per 1M output tokens. Price details and history →
Questions
How much VRAM does DeepSeek V4.1 Flash need?
About 436 GB at 4-bit (Q4_K_M) with an 8K context: 434 GB of weights (estimated), 0.7 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 757 GB.
What is the smallest GPU that runs DeepSeek V4.1 Flash?
No single consumer GPU in our table fits it at Q4_K_M with 8K context.
Can I run DeepSeek V4.1 Flash on a Mac?
Not at Q4_K_M with 8K context on any Mac in our table, even a 512 GB Mac Studio, using the default GPU memory limit.
Similar models
Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.
DeepSeek V4.1 Flash 1017
763B · 436 GB at 4-bitDeepSeek V3.2 Exp
685B · 391 GB at 4-bitDeepSeek V3.2
685B · 379 GB at 4-bitGLM 5
754B · 426 GB at 4-bitGLM 5.1
754B · 434 GB at 4-bitGLM 5.3
753B · 430 GB at 4-bitModel facts
| Hugging Face repo | deepseek-ai/DeepSeek-V4.1-Flash |
| Parameters | 763,205,315,794 (763B) · ~188B active (computed from config) |
| Layers / hidden size | 40 / 5120 |
| Attention | GQA; 64 heads, 1 KV heads, head dim 512 |
| KV cache per token (FP16) | 90.0 KiB |
| Max context | 1,048,576 tokens |
| License | mit |
| Released on HF | Sep 10, 2026 |
| Downloads (30d) / likes | 1,663,303 / 4,986 |
Sources: Hugging Face model API · config.json (deepseek-ai/DeepSeek-V4.1-Flash) · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.