cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition GGUF size and VRAM requirements
bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF is a large language model with 23.57 billion parameters, built on the llama architecture. It is released under the apache-2.0 license and has been downloaded 19,452 times.
To run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF locally at a 4,096-token context, its quantized versions need between 7.68 GB (IQ2_XXS, lowest quality) and 45.5 GB (BF16, highest quality) of memory, weights plus KV cache and a system margin included.
For most users the best balance is IQ2_XXS, needing about 7.68 GB. That means bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF fits entirely in the VRAM of an 8 GB GPU or larger, running fully on the GPU.
GGUF file size and memory by quantization
Compare real GGUF weight sizes, estimated KV cache and total memory for Q4, Q5, Q8 and every quantization published in this repository.
| Quant. | Bits | Quality | Weights | KV | Total | Speed~ | Verdict |
|---|---|---|---|---|---|---|---|
| IQ2_XXS | 2.22 | Very low | 6.1 GB | 0.78 GB | 7.68 GB | 65.6 t/s | Fits in VRAM |
| IQ2_XS | 2.45 | Very low | 6.71 GB | 0.78 GB | 8.29 GB | 7.4 t/s | Offload |
| IQ2_S | 2.54 | Very low | 6.96 GB | 0.78 GB | 8.55 GB | 7.2 t/s | Offload |
| IQ2_M | 2.75 | Low | 7.56 GB | 0.78 GB | 9.14 GB | 6.6 t/s | Offload |
| Q2_K | 3.02 | Low | 8.28 GB | 0.78 GB | 9.86 GB | 6.0 t/s | Offload |
| IQ3_XXS | 3.15 | Low | 8.64 GB | 0.78 GB | 10.22 GB | 5.8 t/s | Offload |
| Q2_K_L | 3.24 | Low | 8.89 GB | 0.78 GB | 10.47 GB | 5.6 t/s | Offload |
| IQ3_XS | 3.36 | Fair | 9.23 GB | 0.78 GB | 10.81 GB | 5.4 t/s | Offload |
| Q3_K_S | 3.53 | Fair | 9.69 GB | 0.78 GB | 11.27 GB | 5.2 t/s | Offload |
| IQ3_M | 3.61 | Fair | 9.92 GB | 0.78 GB | 11.5 GB | 5.0 t/s | Offload |
| Q3_K_M | 3.89 | Fair | 10.69 GB | 0.78 GB | 12.27 GB | 4.7 t/s | Offload |
| Q3_K_L | 4.21 | Good | 11.55 GB | 0.78 GB | 13.13 GB | 4.3 t/s | Offload |
| IQ4_XS | 4.33 | Good | 11.88 GB | 0.78 GB | 13.46 GB | 4.2 t/s | Offload |
| Q3_K_XL | 4.41 | Good | 12.1 GB | 0.78 GB | 13.68 GB | 4.1 t/s | Offload |
| IQ4_NL | 4.57 | Good | 12.54 GB | 0.78 GB | 14.12 GB | 4.0 t/s | Offload |
| Q4_0 | 4.58 | Good | 12.57 GB | 0.78 GB | 14.15 GB | 4.0 t/s | Offload |
| Q4_K_S | 4.6 | Good | 12.62 GB | 0.78 GB | 14.2 GB | 4.0 t/s | Offload |
| Q4_K_M | 4.86 | Good | 13.35 GB | 0.78 GB | 14.93 GB | 3.7 t/s | Offload |
| Q4_K_L | 5.03 | Very good | 13.81 GB | 0.78 GB | 15.39 GB | 3.6 t/s | Offload |
| Q4_1 | 5.05 | Very good | 13.85 GB | 0.78 GB | 15.43 GB | 3.6 t/s | Offload |
| Q5_K_S | 5.53 | Very good | 15.18 GB | 0.78 GB | 16.77 GB | 3.3 t/s | Offload |
| Q5_K_M | 5.69 | Very good | 15.61 GB | 0.78 GB | 17.19 GB | 3.2 t/s | Offload |
| Q5_K_L | 5.83 | Very good | 16.0 GB | 0.78 GB | 17.58 GB | 3.1 t/s | Offload |
| Q6_K | 6.57 | Excellent | 18.02 GB | 0.78 GB | 19.6 GB | 2.8 t/s | Offload |
| Q6_K_L | 6.68 | Excellent | 18.32 GB | 0.78 GB | 19.9 GB | 2.7 t/s | Offload |
| Q8_0 | 8.5 | Excellent | 23.33 GB | 0.78 GB | 24.92 GB | — | Insufficient |
| BF16 | 16.0 | Excellent | 43.92 GB | 0.78 GB | 45.5 GB | — | Insufficient |
KV cache computed from the model's exact architecture. Speed is a rough estimate bounded by memory bandwidth.
Frequently asked questions
How much VRAM do you need to run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF?
You need about 7.68 GB of VRAM to run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF entirely on the GPU using the IQ2_XXS quantization (at a 4,096-token context). Smaller quantizations lower the requirement at the cost of quality.
Can I run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF on an 8 GB GPU?
Yes. With 8 GB of VRAM you can run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF fully on the GPU using IQ2_XXS (about 7.68 GB).
Can I run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF on a 16 GB GPU?
Yes. With 16 GB of VRAM you can run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF fully on the GPU using Q4_1 (about 15.43 GB).
Can I run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF on a 24 GB GPU?
Yes. With 24 GB of VRAM you can run bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF fully on the GPU using Q6_K_L (about 19.9 GB).
What is the best quantization for bartowski/cognitivecomputations_Dolphin-Mistral-24B-Venice-Edition-GGUF?
If memory allows, higher bits-per-weight means better quality. A common sweet spot is a Q4_K_M or Q5_K_M quantization, which keeps most of the quality while roughly halving the memory versus 8-bit. Pick the highest quantization that still fits in your VRAM.