Sold by: NYSgpt
Deployed on AWS
Reranks scientific papers against a query. Beats OpenAI embeddings at finding the right paper first: 93% vs 68% top-1 accuracy.
Overview
A cross-encoder reranker for scientific literature, fine-tuned from Alibaba-NLP/gte-reranker-modernbert-base on peer-review-to-paper pairs. Give it a query and candidate papers from any search system; it returns them ranked by true relevance.
Beats OpenAI embeddings
On a held-out benchmark of 3,157 queries over 1,534 scientific papers (9-way ranking with hard negatives), it picks the correct paper first 93% of the time vs 68% for OpenAI's general embeddings (+25 points), with MRR 95.6% vs 74.5%. It lands the right paper in the top 3 98.2% of the time and cut its base model's error rate by 57%. The benchmark is public: https://huggingface.co/datasets/NYSgpt/44b-eval
Input and output
Input: JSON {"query": "text", "documents": ["text", ...]}. Output: relevance scores and a ranking, highest first.
Runs entirely in your account
CPU inference on ml.m5 and ml.c5 instances. Model card: https://huggingface.co/NYSgpt/44b-reranker
Highlights
- Beats OpenAI embeddings: 93% vs 68% top-1 accuracy at ranking scientific papers
- Right paper in the top 3 98.2% of the time on a held-out benchmark
- CPU inference; your data never leaves your AWS account
Details
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.m5.xlarge Inference (Batch) Recommended | Model inference on the ml.m5.xlarge instance type, batch mode | $3.75 |
ml.m5.xlarge Inference (Real-Time) Recommended | Model inference on the ml.m5.xlarge instance type, real-time mode | $3.75 |
ml.m5.2xlarge Inference (Batch) | Model inference on the ml.m5.2xlarge instance type, batch mode | $3.75 |
ml.m5.2xlarge Inference (Real-Time) | Model inference on the ml.m5.2xlarge instance type, real-time mode | $3.75 |
ml.c5.2xlarge Inference (Real-Time) | Model inference on the ml.c5.2xlarge instance type, real-time mode | $3.75 |
Vendor refund policy
Software fees are billed hourly while your endpoint or batch job runs; stop the endpoint to stop charges. Contact Brendan@nysgpt.com about billing issues within 30 days.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
Initial release.
Additional details
Inputs
- Summary
A query and a list of candidate documents: JSON {"query": "text", "documents": ["text", ...], "top_k": 10}, or {"pairs": [["query", "document"], ...]}.
- Limitations for input type
- Queries and documents are trimmed to the lengths used in training.
- Input MIME type
- application/json
{"query": "CRISPR screens in T cells", "documents": ["A CRISPR knockout screen identifies regulators of T cell exhaustion", "Wastewater surveillance of SARS-CoV-2 variants"]}
https://github.com/nyd-user-1/nysgpt-rerankers-sagemaker/blob/main/data/sample-input.json
Support
Vendor support
Support by email from the publisher (Brendan@nysgpt.com ); typical response within one business day. Usage notebook:
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
A 22.7M-parameter reranker for scientific papers. Beats OpenAI embeddings at top-1 accuracy (87% vs 68%) at a fraction of the size and cost.
Biomedical literature reranker that beats the US National Library of Medicine's own MedCPT reranker on concept search.
OpenAI GPT-5.6 Sol brings OpenAI flagship GPT-5.6 model for advanced reasoning, coding, scientific research, cybersecurity, and agentic workflows to Amazon Bedrock.
Biomedical text embeddings for semantic search and RAG. Fine-tuned from bge-m3 on bioRxiv and medRxiv; outperforms NLM MedCPT on retrieval.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.