List your company

Memory Question Answering on ATM-Bench Hard

74.7QS ScoreGPT-5
32.47643.43854.465.362Mar 2, 2026

Evaluation results

MethodLinks
GPT-5Retriever=Oracle, Reasoning Effort=medium2026.0374.7
Gemini 2.5 ProRetriever=Oracle, Reasoning Effort=medium2026.0364.3
Claude Opus 4.5Retriever=Oracle, Reasoning Effort=medium2026.0362.7
Claude Sonnet 4.5Retriever=Oracle, Reasoning Effort=medium2026.0361.9
Gemini 2.5 FlashRetriever=Oracle, Reasoning Effort=medium2026.0355.6
Qwen3-VL-8BRetriever=Oracle, Reasoning Effort=medium, Variant=Instruct2026.0347.3
Qwen3-VL-2BRetriever=Oracle, Reasoning Effort=medium, Variant=Instruct2026.0334.1