Swift-1.5-Qwen3.8-27B

· llama.cpp · 128K context · Radeon RX 7800 XT · 16 GB

UkisAI's post-training of Qwen3.8-27B for long-horizon agentic and coding work: it spends 58.5% fewer thinking tokens than Swift 1.0 while scoring slightly higher. The GSQ-RCO quants reuse ISTA-DASLab's published per-tensor allocation, refined with a Swift importance matrix, so IQ3_S lands at a KLD of 0.0417.

46 tok/sdecode

229 tok/sprefill

128Kcontext window

Tested on this cardOct 24, 2026

Run it

Follow the steps below to download and run this recipe. Omarchy Local AI lists the recipes it can launch on your hardware. The server then answers on http://localhost:8000/v1.

1 Download the weights

hf download ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF Swift-1.5-Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp.gguf \
  --revision d74895bbe5db4bec1e0024e7cc87d59c02d7631a \
  --local-dir ~/models/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF-d74895bb

2 Start the server

docker run --rm \
  --device /dev/dri \
  -p 8000:8080 \
  --shm-size 8g \
  -v ~/models/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF-d74895bb:/models:ro \
  ghcr.io/ggml-org/llama.cpp@sha256:1086b8815273e566bc0f549bac1d6d2dc469e6409a952cca3f502e53d72877e7 \
  -m /models/Swift-1.5-Qwen3.8-27B-GSQ-RCO-IQ3_S-mtp.gguf \
  --host 0.0.0.0 \
  --port 8080 \
  --ctx-size 131072 \
  --n-gpu-layers 999 \
  -b 2048 \
  -ub 512 \
  -t 4 \
  --parallel 1 \
  --cont-batching \
  --poll 0 \
  --flash-attn on \
  --kv-unified \
  --cache-type-k q5_0 \
  --cache-type-v q4_1 \
  --spec-type draft-mtp \
  --spec-draft-n-max 3 \
  --spec-draft-p-min 0.75 \
  --spec-draft-type-k q4_0 \
  --spec-draft-type-v q4_0 \
  --reasoning auto \
  --reasoning-format deepseek \
  --reasoning-effort xhigh \
  --reasoning-preserve \
  --jinja \
  --no-mmproj \
  --metrics

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 all six checks on a real AMD Radeon RX 7800 XT 16GB (owner: Omarchy 4.0.6, RADV GFX1101) on Oct 24, 2026.

Details

Weights: ukisai/Swift-1.5-Qwen3.8-27B-GSQ-RCO-GGUF @ d74895bbe5 ›
Image: ghcr.io/ggml-org/llama.cpp@sha256:1086b8815273e566bc0f549bac1d6d2dc469e6409a952cca3f502e53d72877e7
Launch: llama.cpp-swift-1.5-qwen3.8-27b-iq3-s-mtp-128k ›
Recipe file: registry/recipes/amd/rx-7800-xt-16gb/swift-1.5-qwen3.8-27b.llama.cpp.128k.json ›
Model released: Sep 16, 2026
Tested: owner, AMD Radeon RX 7800 XT 16GB (owner: Omarchy 4.0.6, RADV GFX1101)