Share your thoughts, 1 month free Claude Pro on usSee more

Coding Accuracy on LiveCodeBench

70Accuracy

RSA

Updated 1mo ago

Evaluation Results

Method	Links
RSA 2026.05		70
EVOLIB 2026.05		70
Best-of-N 2026.05		67.5
DC 2026.05		65
Base 2026.05		58.8
MemQ 2026.05		51.43
ExpRAG 2026.05		51.3
RAG 2026.05		50.48
Self-RAG 2026.05		49.52
MemP 2026.05		49.52
Mem0 2026.05		47.62
MemRL 2026.05		45.71
No Mem. 2026.05		44.76
RecursiveMAS 2026.04		42.9
TextGrad 2026.04		39.8
Dense Rollout Baseline 2026.06		39.7
Sparsity (0.86) 2026.06		39.33
Recursive-TextMAS 2026.04		38.7
Single Agent 2026.04		38.6
Single Agent 2026.04		37.4
Vanilla 2025.11		36.7
kvtc 2025.11		36.5
DMS 2025.11		33.4
kvtc 2025.11		31.6
Mixture-of-Agents (MoA) 2026.04		27
LoopLM 2026.04		24.9
General-Reasoner 2026.06		18.4
DARE + TIES 2026.06		16.63
Vanilla 2025.11		16.4
kvtc 2025.11		16.1
DMS 2025.11		16.1
TIES 2026.06		15.66
RAM 2026.06		14.48
RESMERGE 2026.06		13.5
kvtc 2025.11		13.3
TSV-Merge 2026.06		11.94
Spiral 2026.06		9.98
TA 2026.06		9.59
GRPO 2026.06		5.28
Qwen3-4B 2026.06		4.7