---
library_name: transformers
pipeline_tag: text-generation
license: apache-2.0
base_model: Qwen/Qwen3-8B
tags:
- search-agent
- tool-use
- function-calling
- react
- deep-search
- qwen
---

# SearchQwen3-8B

SearchQwen3-8B is a search-agent model post-trained on environment-aligned multi-hop search tasks and verified ReAct trajectories generated by the EasyDistill2 pipeline. The model is designed for structured `search` / `browse` tool interaction and long-horizon evidence integration.

## Model details

- Base model: `Qwen/Qwen3-8B`
- Parameters: 8.19B
- Training: trajectory SFT followed by process-aware post-training
- Interaction: structured Tool-Call (recommended); Search-R1-style text interaction is also evaluated for Qwen2.5 models
- Companion dataset: [`alibaba-pai/SynSearch-Data`](https://huggingface.co/datasets/alibaba-pai/SynSearch-Data)

## Evaluation results

LLM-judge accuracy (%).

### Search-R1-style text interaction

| Model | Multi-hop QA Avg. | Deep Search Avg. | Overall Avg. |
|---|---:|---:|---:|
| Qwen2.5-7B-Instruct (Base) | 45.28 | 18.65 | 31.96 |
| **SearchQwen2.5-7B** | **52.95** | **26.23** | **39.59** |
| Qwen2.5-3B-Instruct (Base) | 30.12 | 14.95 | 22.54 |
| **SearchQwen2.5-3B** | **39.55** | **21.15** | **30.35** |

### Structured Tool-Call interaction

| Model | Multi-hop QA Avg. | Deep Search Avg. | Overall Avg. |
|---|---:|---:|---:|
| Qwen2.5-7B-Instruct (Base) | 45.23 | 24.35 | 33.33 |
| **SearchQwen2.5-7B** | **55.90** | **33.33** | **44.61** |
| Qwen2.5-3B-Instruct (Base) | 36.10 | 7.05 | 21.60 |
| **SearchQwen2.5-3B** | **48.58** | **21.40** | **35.00** |

### Cross-backbone generalization (Structured Tool-Call)

| Model | Deep Search Avg. | Overall Avg. |
|---|---:|---:|
| Qwen3-8B (Base) | 24.50 | 40.31 |
| **SearchQwen3-8B** | **35.42** | **50.31** |


## Usage

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "alibaba-pai/SearchQwen3-8B"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
)
```

For Qwen3, pass `enable_thinking=False` to `apply_chat_template` when reproducing the non-thinking Tool-Call evaluation setting.

## Verified deployment

The release checkpoint was validated by:

1. loading every safetensors shard;
2. loading the full model on an NVIDIA L20 GPU;
3. serving it with vLLM's OpenAI-compatible API;
4. emitting a structured `search` tool call;
5. consuming the tool response and producing the final answer.

See `SHA256SUMS` for artifact checksums.

## License

This derivative model follows the license included in this repository and the terms of its base model.