Systems Optimization on LLM-SQL
0.7331Final ScoreSwarmResearch
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SwarmResearchagent_framework=Claude Code, backbone=Opus 4.6, budget_per_task=$502026.07 | 0.7331 | — | — | |
| CORALAgent Model=Claude Code + Opus 4.62026.04 | 0.731 | 53.3 | 15 | |
| SOTA2026.07 | 0.731 | — | — | |
| SOTA2026.04 | 0.73 | — | — | |
| EvoXAgent Model=Claude Code + Opus 4.62026.04 | 0.726 | 6 | 83 | |
| EvoXiterations=100, backbone=Opus 4.62026.07 | 0.7253 | — | — | |
| ShinkaEvolveAgent Model=Claude Code + Opus 4.62026.04 | 0.724 | 23.8 | 21 | |
| SEAEVOSHINKABackbone=Gemini-3-Flash2026.04 | 0.72 | — | — | |
| CORALagent_framework=Claude Code, backbone=Opus 4.6, budget_per_task=$502026.07 | 0.7195 | — | — | |
| OpenEvolveAgent Model=Claude Code + Opus 4.62026.04 | 0.716 | 6.7 | 100 | |
| GEPABackbone=Gemini-3-Flash2026.04 | 0.7158 | — | — | |
| OpenEvolveBackbone=Gemini-3-Flash2026.04 | 0.7117 | — | — | |
| SEAEVOOPENBackbone=Gemini-3-Flash2026.04 | 0.7104 | — | — | |
| ShinkaEvolveBackbone=Gemini-3-Flash2026.04 | 0.6998 | — | — | |
| Human/SOTA2026.04 | 0.692 | — | — |