SERA 32B 6bit is an open-source language model by mlx-community. Features: 32b LLM, VRAM: 26.6GB, Context: 40K, License: apache-2.0, Quantized, LLM Explorer Score: 0.23.
| LLM Name | SERA 32B 6bit |
| Repository 🤗 | https://huggingface.co/mlx-community/SERA-32B-6bit |
| Base Model(s) | |
| Model Size | 32b |
| Required VRAM | 26.6 GB |
| Updated | 2026-05-25 |
| Maintainer | mlx-community |
| Model Type | qwen3 |
| Model Files | |
| Quantization Type | 6bit |
| Model Architecture | Qwen3ForCausalLM |
| License | apache-2.0 |
| Context Length | 40960 |
| Model Max Length | 40960 |
| Transformers Version | 4.57.1 |
| Tokenizer Class | Qwen2Tokenizer |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 151936 |
| Errors | replace |
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| ...llcpt E5 Baseline Merged 16bit | 40K / 65.8 GB | 1273 | 0 |
| ... E5 Grad Accum 16 Merged 16bit | 40K / 65.8 GB | 574 | 0 |
| Qwen3 32B Bnb 4bit | 40K / 19.2 GB | 43645 | 7 |
| ...line Merged 16bit Merged 16bit | 40K / 65.8 GB | 23 | 0 |
| Qwen3 32B Unsloth Bnb 4bit | 40K / 39.5 GB | 5822 | 15 |
| Qwen3 32B 4bit | 40K / 18.5 GB | 1319 | 5 |
| Qwen3 32B MLX 8bit | 40K / 33.7 GB | 364 | 11 |
| Qwen3 32B MLX 4bit | 40K / 17.4 GB | 343 | 8 |
| Qwen3 32B MLX 6bit | 40K / 25.5 GB | 79 | 4 |
| Qwen3 32B 8bit | 40K / 34.9 GB | 185 | 1 |
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟