mistralai/Magistral-Small-2506
Full fine-tune (fewer layers) vs mistralai/Mistral-Small-3.1-24B-Instruct-2503. mistralai/Magistral-Small-2506 matches part of mistralai/Mistral-Small-3.1-24B-Instruct-2503 (363 of its 585 tensors; the rest of mistralai/Mistral-Small-3.1-24B-Instruct-2503 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 40 of 40 layers.
Tensors changed: Embeddings 1/1, Norms 45/81, MLP 120/120, Attention 160/160, lm_head 1/1
17 derivatives checked
- lmstudio-community/magistral-small-2506-mlx-bf16: Same weights, repacked
- lmstudio-community/Magistral-Small-2506-MLX-4bit: Quantization
- bartowski/mistralai_Magistral-Small-2506-GGUF: Same weights, converted
- RedHatAI/Magistral-Small-2506-FP8: LoRA-style (fewer layers)
- unsloth/Magistral-Small-2506-GGUF: Same weights, converted
- Mungert/Magistral-Small-2506-abliterated-GGUF: LoRA-style
- mradermacher/Magistral-Small-2506-i1-GGUF: Quantization
- mistralai/Magistral-Small-2506_gguf: Quantization
- jkim96/Magistral-Small-2506-DASHQ-Q2-GGUF: Quantization
- TheDrummer/Cydonia-24B-v3.1-GGUF: Full fine-tune
- mradermacher/Magistral-Small-2506-GGUF: Quantization
- DevQuasar/mistralai.Magistral-Small-2506-GGUF: Same weights, converted
- m-polignano/ANITA-NEXT-24B-Magistral-2506-ITA-GGUF: LoRA-style
- unsloth/Magistral-Small-2506-bnb-4bit: Quantization
- BeaverAI/Cydonia-24B-v3i-GGUF: Full fine-tune
- SicariusSicariiStuff/Impish_Magic_24B: LoRA-style
- TheDrummer/Cydonia-24B-v3.1: Full fine-tune