WisModel¶

OpenSift is powered exclusively by WisModel, a model specifically trained for the two core tasks of the search-verification paradigm.

Overview¶

WisModel is developed by the Fudan NLP Lab and WisPaper.ai, as described in the paper WisPaper: Your AI Scholar Search Engine.

Training approach:

Supervised fine-tuning (SFT) on expert-annotated data
Group Relative Policy Optimization (GRPO)
10 academic disciplines, 2,777 queries, 5,879 criteria

Benchmark: Query Understanding & Criteria Generation¶

WisModel significantly outperforms all baseline models in generating search queries and screening criteria:

Model	Semantic Similarity	ROUGE-1	ROUGE-2	ROUGE-L	BLEU	Length Ratio
Qwen-Max	78.1	43.2	23.1	35.8	11.8	168.9
GPT-4o	91.3	64.0	39.4	52.6	21.5	142.2
GPT-5	87.0	53.8	27.6	41.8	13.2	163.3
GLM-4-Flash	82.2	50.0	25.8	42.1	9.9	167.1
GLM-4.6	84.8	55.5	30.2	44.5	14.4	168.1
DeepSeek-V3.2-Exp	90.2	59.3	32.4	48.0	14.4	153.5
WisModel	94.8	74.9	54.4	67.7	39.8	98.2

Benchmark: Paper-Criteria Matching¶

WisModel achieves 93.70% overall accuracy, surpassing the next best model (Gemini3-Pro, 73.23%) by over 20 percentage points:

Model	Insufficient Info	Reject	Somewhat Support	Support	Overall
GPT-5.1	64.30	63.10	31.40	85.40	70.81
Claude-Sonnet-4.5	46.00	66.50	33.30	87.00	70.62
Qwen3-Max	40.80	72.00	44.20	87.20	72.82
DeepSeek-V3.2	57.90	49.20	45.00	87.00	66.82
Gemini3-Pro	67.40	66.80	15.90	91.10	73.23
WisModel	90.64	94.54	91.82	94.38	93.70

WisModel's biggest advantage

On the hardest category — somewhat support — baseline models struggle at 15.9%–45.0%, while WisModel reaches 91.82%.

Getting Access¶

WisModel is available via the WisPaper API Hub. Contact the team to obtain your API key.

Citation¶

@article{ju2025wispaper,
  title={WisPaper: Your AI Scholar Search Engine},
  author={Li Ju and Jun Zhao and Mingxu Chai and Ziyu Shen and ...},
  journal={arXiv preprint arXiv:2512.06879},
  year={2025}
}