Anonymizer-4B is the most powerful model in the Enchanted anonymizer series, built on the Qwen3-4B base and trained with a combination of supervised fine-tuning and GRPO using GPT-4.1 as a judge to achieve highly accurate, semantically consistent anonymization of personally identifiable information (PII). It provides precise PII replacement by generating semantically similar alternatives that maintain context, scoring 9.55/10 in anonymization quality, and is primarily intended for enterprise or research settings requiring top-quality privacy protection, though it requires higher-end hardware for real-time use.
Model Files
File Name
Quant Type
File Size
Anonymizer-4B.BF16.gguf
BF16
8.05 GB
Anonymizer-4B.F16.gguf
F16
8.05 GB
Anonymizer-4B.F32.gguf
F32
16.1 GB
Anonymizer-4B.Q2_K.gguf
Q2_K
1.67 GB
Anonymizer-4B.Q3_K_L.gguf
Q3_K_L
2.24 GB
Anonymizer-4B.Q3_K_M.gguf
Q3_K_M
2.08 GB
Anonymizer-4B.Q3_K_S.gguf
Q3_K_S
1.89 GB
Anonymizer-4B.Q4_K_M.gguf
Q4_K_M
2.5 GB
Anonymizer-4B.Q4_K_S.gguf
Q4_K_S
2.38 GB
Anonymizer-4B.Q5_K_M.gguf
Q5_K_M
2.89 GB
Anonymizer-4B.Q5_K_S.gguf
Q5_K_S
2.82 GB
Anonymizer-4B.Q6_K.gguf
Q6_K
3.31 GB
Anonymizer-4B.Q8_0.gguf
Q8_0
4.28 GB
Quants Usage
(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)
Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):
