Qwen3.8-27B

EXL3 5 bpw · tabbyapi · 256K context · RTX PRO 4500 Blackwell 32 GB

Alibaba's 27B dense model from August 2026, and the strongest model that fits one consumer card. It thinks before it answers, calls tools reliably and reads images, which makes it the default for coding agents and long projects.

65 tok/sdecode

– tok/sprefill

256Kcontext window

Earlier checkSep 22, 2026

Run it

With Omarchy Local AI it is one button. By hand, it is three steps: the weights, the config, the container. The server then answers on http://localhost:8000/v1.

1 Download the weights

hf download turboderp/Qwen3.8-27B-exl3 \
  --revision f33f26d929e2b20ef21361145d582f5239e3831f \
  --local-dir ~/models/Qwen3.8-27B-exl3-f33f26d9

2 Write the server config

cat > qwen3.8-27b.tabbyapi.256k.yml <<'EOF'
network:
  host: 0.0.0.0
  port: 5000
  disable_auth: true
  disable_fetch_requests: true
  send_tracebacks: false
  api_servers: ["OAI"]
  sse_ping_interval: 15

logging:
  log_prompt: false
  log_generation_params: false
  log_requests: false
  log_chat_completion_requests: false

model:
  model_dir: /workspace/models
  inline_model_loading: false
  use_dummy_models: false
  model_name: Qwen3.8-27B-EXL3-SC5bpw-H6-V6
  backend: exllamav3
  max_seq_len: 262144
  cache_size: 263168
  cache_mode: Q4
  tensor_parallel: false
  gpu_split_auto: true
  autosplit_reserve: [512]
  chunk_size: 2048
  output_chunking: true
  max_batch_size: 4
  vision: true
  vision_offload: false
  reasoning: true
  start_in_reasoning: always
  reasoning_start_token: " thinking"
  reasoning_end_token: "</think>"
  template_vars_default:
    enable_thinking: true
  tool_format: qwen3_coder

draft_model:
  draft_mode: mtp

sampling:
  override_preset:

memory:
  sysmem_recurrent_cache: 4096
  sysmem_kv_cache: 0
  cuda_malloc_async: true
EOF

3 Start the server

docker run --rm \
  --gpus all \
  -p 8000:5000 \
  --shm-size 8g \
  -e NVIDIA_VISIBLE_DEVICES=all \
  -v ~/models/Qwen3.8-27B-exl3-f33f26d9:/workspace/models/Qwen3.8-27B-EXL3-SC5bpw-H6-V6:ro \
  -v $PWD/qwen3.8-27b.tabbyapi.256k.yml:/app/config.yml:ro \
  --entrypoint /opt/venv/bin/python3 \
  ghcr.io/0xsero/tabbyapi-exl3@sha256:0f83e6198dc3be2561652d8df2525a7d1a69733e12fc8e25cbc5bc319a8a3ad1 \
  main.py \
  --config /app/config.yml

What it passed

✓ loadLoads and serves within an hour

✓ chatAnswers a plain question and stops on its own

– reasoningThinks separately and gets 17 × 23 right

– toolsCalls a tool with the right arguments and uses the result

– contextRecalls a code buried in a prompt that fills 85% of the window

– speedDecodes at 15 tok/s or more

Passed the older check (loads and chats); a full six-check run is pending.

Details

Weights: turboderp/Qwen3.8-27B-exl3 @ f33f26d929 ›
Image: ghcr.io/0xsero/tabbyapi-exl3@sha256:0f83e6198dc3be2561652d8df2525a7d1a69733e12fc8e25cbc5bc319a8a3ad1
Engine profile: tabbyapi-qwen3.8-27b-exl3-5bpw-256k ›
Recipe file: registry/recipes/nvidia/rtx-pro-4500-blackwell-32gb/qwen3.8-27b.tabbyapi.256k.json ›
Model released: Aug 5, 2026
Tested: earlier acceptance