XERJ launches Reader-1 document-answering model priced at $0.01 per 1M input tokens

XERJ introduces Reader-1, a neurosymbolic model trained to return the exact source passages that answer a question instead of generating text.

XERJ says each answer is a replayable program with direct traceability to the document, and that Reader-1 is 20-90x faster and 10-1,000x cheaper than alternatives. Developer Santiago (@svpino) reports pricing of $0.01 per 1M input tokens with free output, and quotes the model's description as "not allowed to generate text." The speed and cost comparisons are XERJ's claims, and this report does not include independent benchmarks.