---
title: LingSante-3.1-flash
canonical_url: "https://www.modelscope.cn/models/inclusionAI/LingSante-3.1-flash"
md_url: "https://www.modelscope.cn/models/inclusionAI/LingSante-3.1-flash.md"
repository: inclusionAI/LingSante-3.1-flash
last_updated: 2026-11-22
license: mit
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - bailing_hybrid
architectures:
  - BailingMoeV3ForCausalLM
parameters: 572.1B
tensor_type:
  - F8_E4M3
  - F32
  - BF16
library_name:
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 5
stars: 0
---

# LingSante-3.1-flash

> LingSante-3.1-flash - inclusionAI 在 ModelScope 开源的模型。🤗 Hugging Face &nbsp;&nbsp; | &nbsp;&nbsp;🤖 ModelScope &nbsp;&nbsp; | &nbsp;&nbsp;🔗 Computrix(API) &nbsp;&nbsp;

inclusionAI/LingSante-3.1-flash 是 ModelScope 魔搭社区上的 572.1B 参数text-generation模型，采用 mit 许可。

- **Repository**: inclusionAI/LingSante-3.1-flash
- **License**: mit
- **Tasks**: text-generation
- **Parameters**: 572.1B
- **Downloads**: 5
- **Stars**: 0
- **Last updated**: 2026-11-22

Source: https://www.modelscope.cn/models/inclusionAI/LingSante-3.1-flash

---

<p align="center">
    <img src="https://mdn.alipayobjects.com/huamei_qa8qxu/afts/img/A*4QxcQrBlTiAAAAAAQXAAAAgAemJ7AQ/original" width="100"/>
</p>
<p align="center">🤗 <a href="https://huggingface.co/inclusionAI/LingSante-3.1-flash">Hugging Face</a>&nbsp;&nbsp; | &nbsp;&nbsp;🤖 <a href="https://www.modelscope.cn/models/inclusionAI/LingSante-3.1-flash">ModelScope </a>&nbsp;&nbsp; | &nbsp;&nbsp;🔗 <a href="https://computrix.ai/sg/models/modelservice-1795159585527001414">Computrix(API) </a>&nbsp;&nbsp;</p>

## 
# LingSante-3.1-flash
## Introduction
**LingSante-3.1-flash** is an open-weight Mixture-of-Experts model that combines complex medical reasoning, professional healthcare communication, evidence-based deep research, and long-horizon medical tasks. Built on [Ling-3.1-flash](https://huggingface.co/inclusionAI/Ling-3.1-flash), it has **560 billion total parameters**, activates **25 billion parameters per token**, and supports a **1-million-token context window**.

Real-world medical tasks present challenges that go beyond general intelligence, including incomplete health information, substantial patient heterogeneity, and strict operational and compliance requirements. To address these demands, LingSante-3.1-flash is trained to unify medical reasoning and workflow execution through native medical-mixed pretraining, mid-training, supervised fine-tuning, task-specific reinforcement learning, and multi-teacher on-policy distillation. The model can verify evidence across scenarios, deliver professional judgments, and independently complete multi-stage workflows.

> LingSante-3.1-flash is intended for research and evaluation. It is not a medical device and must not be used as an independent basis for diagnosis, prescriptions, or treatment.
>

<img src="https://modelscope.cn/models/inclusionAI/LingSante-3.1-flash/resolve/master/benches.png" />

## Key Capabilities
+ **Complex Medical Reasoning.** Beyond accurate knowledge understanding, complex medical reasoning requires models to identify the key factors behind clinical decisions, validate supporting evidence, and resolve potential conflicts across evidence sources. To support this, we develop a medical evidence–reasoning alignment framework that strengthens the model’s reasoning foundation through high-quality data synthesis and systematic training. Built on this foundation, LingSante-3.1-flash delivers leading performance among open-source models on complex medical reasoning tasks.
+ **Long-Horizon Medical Tasks.** Long-horizon medical tasks differ fundamentally from those in general domains. They require models to operate across interconnected sub-scenarios, coordinate tools and information within a single workflow, and adapt their plans as new information emerges. To support this, we develop a long-horizon medical task framework that combines realistic interactive environments, diverse tool-use settings, and structured verification. This approach strengthens the model’s ability to execute complex workflows reliably in dynamic medical settings, providing a scalable foundation for training on real-world medical task execution.
+ **Medical Professional Behavior Alignment.** Medical tasks in the real world often involve incomplete information, evolving context, and strict safety expectations. The model must deliver responses that are context-aware, professional, and actionable. To support this, we develop a professional behavior alignment framework that aligns model responses with real clinical scenarios and medical professional standards. LingSante-3.1-flash demonstrates strong performance on benchmarks for medical professional behavior, ranking second overall behind only GPT-6 Astra.



## Model Overview
LingSante-3.1-flash retains the architecture of Ling-3.1-flash.

| Specification | Value |
| --- | --- |
| Architecture | Hybrid-linear MoE |
| Total parameters | 560B |
| Activated parameters per token | 25B |
| Context window | 1M tokens |
| Transformer layers | 63 KDA + 9 Gated MLA (7:1) |
| Dense layers | 4 |
| Routed experts | 512 |
| Shared experts | 1 |
| Activated routed experts per token | 8 |
| Attention heads | 64 |
| Hidden size | 4096 |
| Expert intermediate size | 1280 |
| Dense intermediate size | 10240 |
| Vocabulary size | 157184 |


## Training
General and medical data are used together throughout pretraining, mid-training, and supervised fine-tuning. Medical data provide domain knowledge and clinically grounded training examples, while general data help maintain broad capabilities such as reasoning, coding, search, and tool use.

During post-training, we further improve the model with reinforcement learning for medical reasoning, tool use, search, and professional behavior. These training setups are designed to strengthen not only task performance, but also the quality, reliability, and efficiency of execution under real-world constraints.

Finally, we bring these capabilities together through multi-teacher on-policy distillation, transferring complementary strengths into a single model while preserving strong performance across both medical and general tasks.



## Evaluation
We evaluate LingSante-3.1-flash across a wide range of medical and general benchmarks. The model performs strongly on representative benchmarks for medical knowledge, reasoning, and alignment, while also achieving strong results on internal evaluations for diagnostic understanding, safety, and evidence-based reasoning. Beyond medical QA, LingSante-3.1-flash shows strong performance in more realistic medical settings that require search, verification, and tool use. These results highlight the model’s ability to move beyond answer generation toward evidence-grounded medical problem solving. LingSante-3.1-flash also performs well on benchmarks that reflect professional medical workflows and long-horizon execution. Overall, the results show that LingSante-3.1-flash combines strong medical reasoning, professional behavior, and workflow execution in a single model.

<img src="https://modelscope.cn/models/inclusionAI/LingSante-3.1-flash/resolve/master/benches-charts.png" />

## Quickstart
### SGLang (Recommended)
The day-0 SGLang image includes LingSante-3.1-flash support, MTP (NEXTN) speculative decoding, and Hopper decode optimizations:

```bash
docker pull lmsysorg/sglang:dev-ling-3.1-flash
```

#### Run Inference
**Server** (8 × H200, TP8 + EP8):

```bash
docker run --gpus all --network host --ipc=host --rm \
  lmsysorg/sglang:dev-ling-3.1-flash \
  python3 -m sglang.launch_server \
    --model-path inclusionAI/LingSante-3.1-flash \
    --tp 8 --ep 8 \
    --speculative-algorithm NEXTN \
    --host 0.0.0.0 --port 30000
```

On 4 × GB300, use `--tp 4 --ep 4` instead. Hopper decode optimizations (optional, ~10% faster batch-1 decode) can be enabled with `SGLANG_ENABLE_FP8_BLOCK_GEMV=1` and `SGLANG_ENABLE_FUSED_KDA_OUTPUT_QUANT=1`.

**Client**

```bash
curl -s http://localhost:30000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "LingSante-3.1-flash",
       "messages": [{"role": "user", "content": "What is the capital of France?"}],
       "chat_template_kwargs": {"enable_thinking": true},
       "stream": true,
       "temperature": 1.0,
       "top_k": 20,
       "top_p": 0.95}'
```

Verified recipes for H200, GB300, H800, H100, and B200, plus HiCache (Mooncake L3), reasoning on/off, tool calling, and measured accuracy/speed numbers, are in the [SGLang Ling-3.1-flash cookbook](https://docs.sglang.io/cookbook/autoregressive/InclusionAI/Ling-3.1-flash).

### vLLM
#### Environment Preparation
```bash
pip install uv

uv venv ~/my_ling_env

source ~/my_ling_env/bin/activate

git clone -b vllm https://github.com/vllm-project/vllm.git

cd vllm

VLLM_USE_PRECOMPILED=1 uv pip install --editable . --torch-backend=auto
```

#### Run Inference
**Server**

```bash
vllm serve "$MODEL_PATH" \
    --port "$PORT" \
    --trust-remote-code \
    --served-model-name auto \
    --tensor-parallel-size 8 \
    --gpu-memory-utilization 0.85 \
    --enable-prefix-caching \
    --enable-expert-parallel \
    --mamba-cache-mode align \
    --enable-auto-tool-choice \
    --tool-call-parser ling3 \
    --reasoning-parser ling3 \
    --speculative-config '{"method":"mtp","num_speculative_tokens":3}'
```

**Client**  
It is recommended to use the sampling parameters (`temperature=1.0`, `top_p=0.95`, `top_k=20`) and set `enable_thinking` to `true` for better performance.

```bash
curl -s http://${MASTER_IP}:${PORT}/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model": "auto",
       "messages": [{"role": "user", "content": "What is the capital of France?"}],
       "chat_template_kwargs": {"enable_thinking": true},
       "stream": true,
       "temperature": 1.0,
       "top_k": 20,
       "top_p": 0.95}'
```

## Intended Use
+ Medical and healthcare research, including literature analysis and evidence synthesis.
+ Medical question answering and clinical document analysis with qualified human review.
+ Structured information extraction and retrieval-augmented healthcare applications.
+ Development and evaluation of medical, research, and general-purpose tool-using agents.
+ Research on professional behavior, multi-agent search, and cross-task capability integration.

## Limitations and Responsible Use
+ Medical outputs may be incorrect, incomplete, outdated, or unsupported by the cited evidence. Verify consequential claims against current authoritative sources.
+ The model does not replace qualified healthcare professionals. It must not independently diagnose, prescribe, determine treatment, or make clinical triage decisions.
+ Benchmark and simulated-environment results do not establish clinical effectiveness. Performance may vary across specialties, languages, populations, institutions, and local resource conditions.
+ Long-context support does not guarantee complete retrieval of every relevant fact. Check source attribution, omissions, and contradictions when processing clinical records or literature.
+ Protect personal health information and comply with applicable privacy, security, healthcare, and data-governance requirements.
+ Use least-privilege tool access, auditable execution, and human approval for consequential actions. Search results and tool outputs can contain errors or malicious instructions.

## Model Lineage
+ Base model: [Ling-3.1-flash](https://huggingface.co/inclusionAI/Ling-3.1-flash).
+ Previous healthcare release: [Ling-3.0-flash-Sante](https://huggingface.co/inclusionAI/Ling-3.0-flash-Sante), with 124B total and 5.1B active parameters per token.
