Instructions to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Use Docker

docker model run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Ollama
How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with Ollama:
```
ollama run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
```
Unsloth Desktop

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with Docker Model Runner:
```
docker model run hf.co/ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
```

Lemonade

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Swift-1.5-Qwen3.8-Flash-Next-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Website • Learn more • GGUF • GSQ-RCO GGUF • Evaluation • Enterprise licensing

Swift 1.5 Qwen3.8-Flash-Next

GGUF quantizations. Derived directly from Swift 1.5 Qwen3.8-Flash-Next with llama.cpp. Run a chosen quantization tier with a current llama.cpp-compatible runtime such as llama-server.

Swift 1.5 Qwen3.8-Flash-Next is UkisAI's reasoning-efficient derivative of Qwen3.8-Flash-Next. It uses 63.4% fewer thinking tokens, with a 1.8x speed up while keeping the accuracy loss <1% vs base on xhigh.

Demo

We gave base Qwen3.8-Flash-Next and Swift 1.5 Qwen3.8-Flash-Next the same prompt:

Create a 3D endless runner that has the fast, playful feel of Subway Surfers, but make the world and characters your own. I want to run through a lively place, dodge things, collect rewards, and feel the pace build the longer I survive. Make it fun to control and visually memorable. Use your judgment for the setting, mechanics, and little details that make it feel like a real game. Build it so I can launch and play it locally, then tell me how to run it.

Try the game yourself here: https://ukisai.com/swift-games/flash-next

Base Qwen3.8-Flash-Next took 8 minutes 52 seconds to build its game. Swift 1.5 took 4 minutes 56 seconds.

Training approach

We made Swift Flash Next efficient by figuring out which tokens were linked to pathological overthinking and penalizing them without "attacking" the reasoning length directly then regained the accuracy with RL and OPD, leading to "compressed" token usage while maintaining accuracy.

Swift 1.5 produces shorter reasoning traces. In our testing, we also observe fewer overthinking errors.

This release also features our previously mentioned post-training methods adapted specifically for coding and long-horizon agent work such as personal agents, terminal use and software engineering.

Our training data is viewable here: https://huggingface.co/datasets/ukisai/Qwen3.8-27B-multi-turn-agent-sft albeit it is not used out of the box, but rather re-sampled, turned into proper RL environments etc.

Evaluation

All scores compare the Qwen3.8-Flash-Next BF16 base with the Swift 1.5 BF16 checkpoint. Token columns report thinking tokens, except Terminal-Bench 2.1, which reports total generated output tokens.

Benchmark	Score		Mean tokens			Median tokens
Benchmark	Base	Swift 1.5	Base	Swift 1.5	Reduction	Reduction
General reasoning
GPQA-Diamond	89.80%	89.60%	17,683	7,823	↓ 55.8%	↓ 63.4%
MMLU-Pro	87.75%	87.20%	3,528	1,519	↓ 57.0%	↓ 24.0%
C-Eval	93.27%	93.60%	1,048	586	↓ 44.1%	↓ 7.1%
IFBench	73.20%	70.13%	8,310	4,411	↓ 46.9%	↓ 55.6%
Mathematics
AIME 2026	98.67%	96.67%	23,015	15,806	↓ 31.3%	↓ 51.1%
HMMT (Nov 2025)	98.00%	97.33%	25,487	16,530	↓ 35.1%	↓ 54.7%
Multimodal
ERQA	70.80%	69.30%	4,036	1,788	↓ 55.7%	↓ 47.7%
Coding
LiveCodeBench v6	88.40%	90.39%	17,833	9,849	↓ 44.8%	↓ 52.0%
Agentic coding
Terminal-Bench 2.1	67.64%	69.66%	40,591	45,428	↑ 11.9%	↓ 17.9%

How to reproduce

Serving: BF16 · Qwen3 reasoning parser · context 262,144 · thinking xhigh · MTP disabled.
Sampling: temperature 1.0 · top_p 0.95 · top_k 20 · min_p 0 · presence_penalty 0 · repetition_penalty 1.
Benchmarks: five seeds (0–4) for the seeded question benchmarks; Terminal-Bench 2.1 uses five attempts per task. LiveCodeBench is full release v6 mean pass@1 over seeds.
Terminal-Bench 2.1: Harbor 0.20.0 / Terminus-2 2.0.0, pinned 89-task dataset, JSON parser, interleaved thinking, temperature 1, top_p 1, 131,072-token server context, 3,600-second LLM call timeout and native per-task limits. Swift used concurrency 8; its context-recovery fix was applied during the run.

Benchmark	Output cap
GPQA-Diamond	100,000
MMLU-Pro	100,000
C-Eval	16,384
IFBench	81,920
AIME 2026	250,000
HMMT Nov 2025	250,000
ERQA	100,000
LiveCodeBench v6	100,000

Efficiency across reasoning efforts

GPQA-Diamond at each reasoning_effort setting, Swift 1.5 against the base at the same setting:

Reasoning effort	Score		Mean tokens			Median tokens
Reasoning effort	Base	Swift 1.5	Base	Swift 1.5	Reduction	Reduction
Xhigh	89.80%	89.60%	17,683	7,823	↓ 55.8%	↓ 63.4%
Medium	86.36%	83.74%	4,157	2,483	↓ 40.3%	↓ 25.1%
Low	87.17%	84.75%	3,966	2,645	↓ 33.3%	↓ 19.7%

At xhigh, Swift 1.5 trails base by 0.20 percentage points while using 55.8% fewer mean and 63.4% fewer median thinking tokens. Medium and low save tokens but also lose 2.62 and 2.42 percentage points respectively.

Quantized models

Format	Repository	Runtime
AWQ INT4 (W4A16)	Swift-1.5-Qwen3.8-Flash-Next-W4A16-AWQ	vLLM (`compressed-tensors`)
AutoRound INT4 (W4A16)	Swift-1.5-Qwen3.8-Flash-Next-W4A16-AutoRound	vLLM (`auto-round`)
NVFP4	Swift-1.5-Qwen3.8-Flash-Next-NVFP4	NVIDIA Blackwell
GGUF	Swift-1.5-Qwen3.8-Flash-Next-GGUF	llama.cpp
GSQ-RCO GGUF (compact 2–3 bit)	Swift-1.5-Qwen3.8-Flash-Next-GSQ-RCO-GGUF	llama.cpp

GGUF quantizations

File	Size	KLD wikitext @512	KLD wikitext @32k	99% KLD @32k	Top-p @32k
Q8_0	188.3 GB	0.0297	0.0220	0.230	94.91%
Q6_K	168.1 GB	0.0324	0.0252	0.256	94.34%
Q5_K_L	151.1 GB	0.0788	0.0575	0.610	91.61%
Q4_K_L	139.3 GB	0.1051	0.0740	0.791	90.14%
Q5_K_M	134.7 GB	0.1641	0.1386	1.654	87.34%
Q5_K_S	128.1 GB	0.1777	0.1482	1.763	86.92%
Q4_K_M	119.6 GB	0.1512	0.1100	1.254	88.20%
Q3_K_XL	119.3 GB	0.3718	0.3250	3.303	79.79%
Q4_K_S	113.1 GB	0.1779	0.1311	1.420	87.01%
Q4_1	111.3 GB	0.2532	0.1812	2.048	85.07%
Q2_K_L	107.1 GB	0.5425	0.3867	3.654	77.28%
Q4_0	100.6 GB	0.2524	0.2593	2.798	82.53%
IQ4_NL	100.3 GB	0.1261	0.0926	1.010	88.96%
IQ4_XS	97.7 GB	0.1305	0.0929	1.010	88.91%
IQ3_M	93.2 GB	0.2005	0.1482	1.605	86.22%
Q3_K_L	93.2 GB	0.3842	0.3300	3.259	79.44%
Q3_K_M	92.0 GB	0.3210	0.2260	2.401	82.73%
IQ3_XS	91.9 GB	0.2211	0.1644	1.738	85.26%
Q3_K_S	89.4 GB	0.3253	0.2284	2.399	82.59%
IQ3_XXS	88.0 GB	0.2868	0.2185	2.350	83.30%
Q2_K	80.9 GB	0.5703	0.4091	3.832	76.66%
IQ2_M	80.4 GB	0.3769	0.2759	2.708	80.76%
IQ2_S	77.8 GB	0.4846	0.4331	4.196	76.47%
IQ2_XS	77.7 GB	0.5136	0.4033	3.798	76.93%
IQ2_XXS	75.2 GB	0.6052	0.4472	4.002	75.58%
IQ1_M	72.0 GB	0.9067	0.6928	5.392	69.57%
IQ1_S	70.1 GB	1.1601	0.9343	6.726	65.15%

Mean KL divergence against the Swift 1.5 BF16 source, lower is better. wikitext @512 is wikitext-2 test, 100 windows of 512 tokens. wikitext @32k is wikitext-2 test, 9 windows of 32,768 tokens, scoring the second half of each window, so every scored token sees at least 16k tokens of context. 99% KLD is the 99th-percentile divergence on the same 32k run. Top-p is top-token agreement with BF16 on the 32k run.

Each tier is in its own folder, split into several parts. llama.cpp loads every part from the first one:

llama-server -hf ukisai/Swift-1.5-Qwen3.8-Flash-Next-GGUF:Q4_K_M
# or, after downloading a folder:
llama-server -m Q4_K_M/Swift-1.5-Qwen3.8-Flash-Next-Q4_K_M-00001-of-00003.gguf

Use case	Pick
128 GB unified memory, everyday use	IQ4_XS
Long agentic runs, strict tool-call formatting	Q6_K or higher
Maximum fidelity	Q8_0

Recipe

All 27 tiers were built with llama.cpp commit d2e5458 from a BF16 conversion of the published Swift 1.5 safetensors. They use bartowski's exact per-tensor type map read from bartowski/Qwen3.8-Flash-Next-GGUF, with an importance matrix computed on this model from his calibration-v6 corpus (-c 512, 583 chunks, --parse-special). Every file was checked against BF16 on the harness above.

License and access

Swift 1.5 Qwen3.8-Flash-Next is a derivative of Qwen3.8-Flash-Next (Copyright (c) 2026 Qwen, Qwen Community License 1.0). UkisAI's contribution, including the adapted weights, is licensed under the Swift Open License v1.0. See NOTICE for the change notice and attribution details.

Personal, research, educational, evaluation, and commercial use of the Swift contribution are free for individuals and organizations with gross annual revenue, including affiliates, of up to US$1,000,000. Above that threshold, commercial use requires a separate Swift Enterprise License. Contact UkisAI for terms.

The base model's terms still apply to it. Under the Qwen Community License 1.0, organizations that run a Model-as-a-Service or AI Work Assistant business need a separate license from Qwen before any commercial use, and products above 100 million monthly active users or US$20 million monthly revenue must prominently display the model name. Nothing in the Swift Open License limits your rights in Qwen3.8-Flash-Next itself under the Qwen Community License.

Citation

@misc{swift-qwen3.8-flash-next,
  title  = {Swift 1.5 Qwen3.8-Flash-Next},
  author = {UkisAI},
  year   = {2026},
  url    = {https://huggingface.co/ukisai/Swift-Qwen3.8-Flash-Next}
}

Acknowledgements

We acknowledge the NVIDIA Innovation Lab, Amazon Web Services, and Google Cloud for providing compute credits and infrastructure support for Swift's development, training, and evaluation.