BAAI/AREX-Turbo

Full fine-tune vs Qwen/Qwen3.5-4B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/4, MLP 96/196, Attention 196/320, Norms 81/203

4 derivatives checked

• bartowski/BAAI_AREX-Turbo-GGUF: Same weights, converted
• CloudGoat/Mephisto-1.0-4B: Full fine-tune (fewer layers)
• CloudGoat/Mephisto-1.5-4B: Full fine-tune (fewer layers)
• DareModels/dare4b: Full fine-tune

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse