DailyReport (Claude Code) - Factuality: leaderboard
DailyReport (2026) tests search agents on open-ended daily information-seeking tasks about trending topics; each task is split into subtasks with cascade rubrics for instruction following, factuality and rationality, judged by Gemini 3 Flash. This board is the percent of extracted factual claims verified correct by web search, counted where the instruction-following rubric holds, for LLMs running inside the Claude Code agent with search tools over MCP; higher is better.
Metric: Factuality (0-1 scaled to %), the share of extracted objective claims verified correct by web search, counted only where the instruction-following rubric holds; open-ended daily search tasks on trending topics decomposed into subtasks with cascade rubrics, judged by Gemini 3 Flash; LLMs as backends of the Claude Code agent with Serper search and Jina Reader over MCP, extended thinking enabled; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 (Medium) | 81.3 | #123 (GPT-5.4) |
| 2 | GLM-5 | 76.7 | #260 |
| 3 | Claude Opus 4.6 (Thinking) | 75.6 | #102 (Claude Opus 4.6) |
| 4 | Kimi K2.5 | 71.8 | #256 |
| 5 | Gemini 3.1 Pro (Preview) | 68.4 | #94 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=dailyreport-claude-code-factuality · How It Works · Data refreshed daily, snapshot 2026-11-13.