A 76-layer, 31.9B-parameter model created by transplanting the three most-drifted organs from Qwen3.6-27B into Qwen3.8-27B. See the bf16 model card for the full organ selection methodology.

Key Stats

Parameters: ~31.9B
Layers: 76
Precision: 8-bit (mlx-lm quantized)
MTP Compatible: Accepts the parent Qwen3.8 MTP drafter for speculative decoding

Usage (MLX)

bash

# Standard generation
mlx_vlm.generate \
  --model StargazerLabs/Qwen3.8-32B-Jumbo-8bit \
  --prompt "Your prompt here" --max-tokens 2048

# With MTP speculative decoding
mlx_vlm.generate \
  --model StargazerLabs/Qwen3.8-32B-Jumbo-8bit \
  --draft-model mlx-community/Qwen3.8-27B-MTP-8bit \
  --prompt "Your prompt here" --max-tokens 2048

StargazerLabs/Qwen3.8-32B-Jumbo-8bit

Qwen3.8-32B-Jumbo (8-bit)

Key Stats

Usage (MLX)

Related

StargazerLabs/
Qwen3.8-32B-Jumbo-8bit