OR Batch Inference Models
OpenRouter Batch Inference Models Every model with a :batch id on OpenRouter on 2026-10-31 (71 models), with batch and sync prices, the providers that serve batch, and a price-tier label. Source: public https://openrouter.ai/api/v1/models and /api/v1/models/{id}/endpoints, fetched 2026-10-31T14:52:01Z. No key needed. Prices: US$ per million tokens, base tier, from pricing.prompt / pricing.completion × 1e6. Each snapshot is a split named snapshot_YYYY_MM_DD. New snapshots are… See the full description on the dataset page: https://huggingface.co/datasets/danielrosehill/OR-Batch-Inference-Models.
About This Project
This project offers a periodically updated repository of OpenRouter models that support batch inference. By aggregating data from the OpenRouter API, it provides a standardized view of batch and synchronous pricing, associated service providers, and technical model attributes for 71 distinct model identifiers. The dataset is designed to assist developers and researchers in tracking the evolving cost structures and availability of LLM batch processing.
The dataset includes enriched metadata such as price-tier labels, input-to-output ratios, and endpoint stability assessments, allowing users to analyze cost trends and model version reliability. It is particularly useful for those building cost-optimized machine learning pipelines or conducting comparative analysis of LLM inference economics across various providers.