VibeSearchBench - Daily: leaderboard

Metric: Triplet F1 (%; an LLM judge matches the knowledge graph the agent extracts after a multi-turn session with a progressive-disclosure user simulator against the annotated ground-truth graph; OpenClaw agent harness; 100 everyday VibeSearch-Daily tasks; mean of 3 runs per task). Source: arxiv.org. Saturation forecast: Around 2029. 7 models tracked.

Top models

#ModelScore
1Claude Opus 4.628.04
2DeepSeek V4 Pro26.16
3Gemini 3.1 Pro (Preview)25.54
4Kimi K2.624.93
5Seed 2.0 Pro24.64
6Qwen 3.5 397B A17B21.77
7GPT-5.421.05

Interactive version: theaggregate.ai/benchmark?slug=vibesearchbench-daily · How It Works · Data refreshed daily, snapshot 2026-09-26.