List your company
ResearchFormerly WizWand

Average QA and Reasoning on NQ, TriviaQA, HotpotQA, and GSM8K

72.5Average AccuracyKite-best
58.56462.18265.869.418Sep 19, 2025
Updated 5mo ago

Evaluation results

MethodLinks
Kite-bestModel=GPT-4.1, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0972.5
DPP-bestModel=GPT-4.1, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0971.7
KNN-bestModel=GPT-4.1, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0971.5
BM25Model=GPT-4.1, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0971.2
Kite-bestModel=GPT-5.2, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0971
DPP-bestModel=GPT-5.2, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0970.1
KNN-bestModel=GPT-5.2, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0969.8
BM25Model=GPT-5.2, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0968.2
Kite-bestModel=GPT-4o-mini, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0965.3
KNN-bestModel=GPT-4o-mini, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0965.1
DPP-bestModel=GPT-4o-mini, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0965.1
BM25Model=GPT-4o-mini, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0964.1
Kite-bestModel=Claude-Haiku-4.5, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0962.1
DPP-bestModel=Claude-Haiku-4.5, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0961
KNN-bestModel=Claude-Haiku-4.5, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0960.3
BM25Model=Claude-Haiku-4.5, Exemplar budget (r)=5, Embedding Encoder=OpenAI large text embeddings2025.0959.1