List your company

Multi-domain knowledge reasoning on HLE 500-question ablation

57.3Success Rate (Last)MemRL
17.88428.11738.3548.583Mar 31, 2026

Evaluation results

MethodLinks
MemRLBackbone=Gemini-3-pro2026.0357.361.3
MemPBackbone=Gemini-3-pro2026.0352.858.2
Mem0Backbone=Gemini-3-pro2026.0351.256
RAGBackbone=Gemini-3-pro2026.035054.8
Self-RAGBackbone=Gemini-3-pro2026.0348.854.8
APEX-EM: Entity graph (A3, E10)Backbone=Claude Opus 4.52026.034853.3
APEX-EM: Full memory (A2, E10)Backbone=Claude Opus 4.52026.0346.852.3
APEX-EM: Semantic only (A1, E10)Backbone=Claude Opus 4.52026.0345.652.3
APEX-EM: Judge + iteration (A5, E10)Backbone=Claude Opus 4.52026.0340.452.9
No MemoryBackbone=Gemini-3-pro2026.0335.7—
APEX-EM: No Memory (A0)Backbone=Claude Opus 4.52026.0325.2—
APEX-EM: Memory, no judge (A4, E10)Backbone=Claude Opus 4.52026.0319.447.9
Pass@10Backbone=Gemini-3-pro2026.03—52.4
ReflexionBackbone=Gemini-3-pro2026.03—53