eaddario
/

Dolphin3.0-Mistral-24B-GGUF

@@ -69,7 +69,7 @@ By producing imatrix optimized quantized models, we can maintain inference effic
 | [Dolphin3.0-Mistral-24B-Q8_0](./Dolphin3.0-Mistral-24B-Q8_0.gguf)     | Q8_0       | 25.10G | 7.668204 ±0.052714  | 99.94%               | 0.001478 ±0.000032 | Extremely high quality, generally unneeded but max available quant             |
 | [Dolphin3.0-Mistral-24B-Q6_K](./Dolphin3.0-Mistral-24B-Q6_K.gguf)     | Q6_K       | 19.30G | 7.697080 ±0.053150  | 99.90%               | 0.003182 ±0.000018 | Very high quality, near perfect, *recommended*                                 |
 | [Dolphin3.0-Mistral-24B-Q5_K_M](./Dolphin3.0-Mistral-24B-Q5_K_M.gguf) | Q5_K_M     | 16.80G | 7.713033 ±0.053233  | 99.84%               | 0.006054 ±0.000033 | High quality                                                                   |
-| [Dolphin3.0-Mistral-24B-Q5_K_S](./Dolphin3.0-Mistral-24B-Q5_K_S.gguf) | Q5_K_S     | 13.50G | 7.731911 ±0.053337  | 99.83%               | 0.006765 ±0.000058 | High quality, *recommended*                                                    |
 | [Dolphin3.0-Mistral-24B-IQ4_NL](./Dolphin3.0-Mistral-24B-IQ4_NL.gguf) | IQ4_NL     | 13.50G | 7.840359 ±0.054591  | 99.59%               | 0.018718 ±0.000116 | Good quality, new method (super-blocks with 256 weights), *recommended*        |
 | [Dolphin3.0-Mistral-24B-Q4_K_M](./Dolphin3.0-Mistral-24B-Q4_K_M.gguf) | Q4_K_M     | 14.30G | 7.815683 ±0.054370  | 99.64%               | 0.015912 ±0.000100 | Good quality, default size for must use cases, *recommended*                   |
 | [Dolphin3.0-Mistral-24B-Q4_K_S](./Dolphin3.0-Mistral-24B-Q4_K_S.gguf) | Q4_K_S     | 13.50G | 7.853283 ±0.054726  | 99.59%               | 0.018656 ±0.000110 | Good quality, best choice in the Q4 series if RAM is scarce, *recommended*     |

 | [Dolphin3.0-Mistral-24B-Q8_0](./Dolphin3.0-Mistral-24B-Q8_0.gguf)     | Q8_0       | 25.10G | 7.668204 ±0.052714  | 99.94%               | 0.001478 ±0.000032 | Extremely high quality, generally unneeded but max available quant             |
 | [Dolphin3.0-Mistral-24B-Q6_K](./Dolphin3.0-Mistral-24B-Q6_K.gguf)     | Q6_K       | 19.30G | 7.697080 ±0.053150  | 99.90%               | 0.003182 ±0.000018 | Very high quality, near perfect, *recommended*                                 |
 | [Dolphin3.0-Mistral-24B-Q5_K_M](./Dolphin3.0-Mistral-24B-Q5_K_M.gguf) | Q5_K_M     | 16.80G | 7.713033 ±0.053233  | 99.84%               | 0.006054 ±0.000033 | High quality                                                                   |
+| [Dolphin3.0-Mistral-24B-Q5_K_S](./Dolphin3.0-Mistral-24B-Q5_K_S.gguf) | Q5_K_S     | 16.30G | 7.731911 ±0.053337  | 99.83%               | 0.006765 ±0.000058 | High quality, *recommended*                                                    |
 | [Dolphin3.0-Mistral-24B-IQ4_NL](./Dolphin3.0-Mistral-24B-IQ4_NL.gguf) | IQ4_NL     | 13.50G | 7.840359 ±0.054591  | 99.59%               | 0.018718 ±0.000116 | Good quality, new method (super-blocks with 256 weights), *recommended*        |
 | [Dolphin3.0-Mistral-24B-Q4_K_M](./Dolphin3.0-Mistral-24B-Q4_K_M.gguf) | Q4_K_M     | 14.30G | 7.815683 ±0.054370  | 99.64%               | 0.015912 ±0.000100 | Good quality, default size for must use cases, *recommended*                   |
 | [Dolphin3.0-Mistral-24B-Q4_K_S](./Dolphin3.0-Mistral-24B-Q4_K_S.gguf) | Q4_K_S     | 13.50G | 7.853283 ±0.054726  | 99.59%               | 0.018656 ±0.000110 | Good quality, best choice in the Q4 series if RAM is scarce, *recommended*     |