Instructions to use MitakaKuma/BaT-Qwen3.5-9B-0815 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use MitakaKuma/BaT-Qwen3.5-9B-0815 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="MitakaKuma/BaT-Qwen3.5-9B-0815")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
pipe(text=messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("MitakaKuma/BaT-Qwen3.5-9B-0815")
model = AutoModelForMultimodalLM.from_pretrained("MitakaKuma/BaT-Qwen3.5-9B-0815", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"},
            {"type": "text", "text": "What animal is on the candy?"}
        ]
    },
]
inputs = processor.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=256)
print(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use MitakaKuma/BaT-Qwen3.5-9B-0815 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "MitakaKuma/BaT-Qwen3.5-9B-0815"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MitakaKuma/BaT-Qwen3.5-9B-0815",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/MitakaKuma/BaT-Qwen3.5-9B-0815

SGLang

How to use MitakaKuma/BaT-Qwen3.5-9B-0815 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "MitakaKuma/BaT-Qwen3.5-9B-0815" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MitakaKuma/BaT-Qwen3.5-9B-0815",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "MitakaKuma/BaT-Qwen3.5-9B-0815" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "MitakaKuma/BaT-Qwen3.5-9B-0815",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use MitakaKuma/BaT-Qwen3.5-9B-0815 with Docker Model Runner:
```
docker model run hf.co/MitakaKuma/BaT-Qwen3.5-9B-0815
```

Request access to BaT

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Please provide your affiliation and a brief description of your research. Access requests are reviewed manually.

BaT logo

BaT-Qwen3.5-9B-0815

Paper · Website · Code · Data

BaT develops agentic workflow skills through stage-level feedback, building on Qwen3.5-9B.

Training scope: BaT post-training data covers agentic skills only, not professional medical reasoning or medical knowledge. Benchmark performance does not establish clinical expertise. Research use only; not for diagnosis or treatment.

Agent comparisons

Original paper results; checkpoint correspondence remains unverified. Scores: 0–100, higher is better.

BaT uses OpenHands; other systems use the Codex or Claude Code harness shown in the legend. Comparisons follow each benchmark's scorer; MedXpertQA-Text evaluates text responses.

Agentic skills and small-model comparisons

Agentic, task, and overall scores across training methods

Policy comparisons use the default OpenHands runner. Overall = (Task + Agentic) / 2. These differ from full-agent results above.

Policy scores

Model / training	Overall	Agentic	Task
Qwen3.5-9B (base)	19.9	11.3	28.4
Qwen3.5-9B + SFT	12.9	21.6	4.1
Qwen3.5-9B + GRPO	31.9	39.2	24.5
BaT-9B (BiCuRL)	53.4	64.7	42.1

Evaluation settings

Qwen / BaT: default Qwen thinking enabled, xhigh effort, 64K output tokens per request, 200K context window.
Other agents: xhigh effort; Codex with 258K context and Claude Code with 1M context.
Harness freeze: August 1, 2026. Latest stable releases by that date: Codex 0.146.0 and Claude Code 2.1.220. These are release-date references; exact installed builds require run-log verification.

External benchmarks

Direct model API evaluation, compared only with the corresponding base model under the same protocol. Interactive tasks retain benchmark tools and environments.

Benchmark	Qwen3.5-9B (base)	BaT-9B	Change (pp)
AIME 2025	88.3	83.6	-4.7
AIME 2026	92.5	86.7	-5.8
GPQA-Diamond	81.7	78.3	-3.4
τ²-Bench	79.1	84.5	+5.4
BFCL-Parity	79.4	68.3	-11.1
GAIA	24.9	21.2	-3.7
SWE-bench Verified	49.0	51.8	+2.8
Terminal Bench 2.0	19.9	21.3	+1.4

Citation

@article{liu2026bat,
  title = {BaT: Towards Self-Evolving Medical Research Agent with Stage Rubrics},
  author = {Liu, Junqi and He, Yufan and He, Yexiao and Guo, Pengfei and Yang, Dong and Myronenko, Andriy and Zhao, Can and Ye, Hanrong and Qi, Tianhao and Zhou, Yuyin and Xu, Daguang and Tang, Yucheng},
  journal = {arXiv preprint arXiv:2608.16211},
  year = {2026},
  url = {https://arxiv.org/abs/2608.16211}
}