SourceVerify Research

Hallucination Calculator

Predict how well an LLM can generate scholarly references for any research topic.

Architecture type

Model size (active parameters)~70B

Gemma 4B

Llama 8B

Mistral 24B

Llama 70B

Mistral 250B

Llama 405B

GPT-5

8B70B405B3T

Try an example

Have references to check, but don't have a lot of time? SourceVerify.ai

Trust Every Reference