DeepWeb-Bench - Calibration: leaderboard

The calibration capability family of DeepWeb-Bench (2026), a deep-research benchmark that demands massive cross-source web evidence and long-horizon derivation, scored cell by cell. The score is the mean cell score (%) over the calibration cells for agents run in Claude Code or Codex CLI with the search and browsing tools the benchmark provides; higher is better.

Metric: Mean cell score (%) over the Calibration capability family (calibrated answers), DeepWeb-Bench deep-research tasks answered cell by cell with the benchmark's web search, page visit and PDF tools (native browsing disabled), runs hosted in Claude Code CLI or Codex CLI at default settings; higher is better. Source: arxiv.org. Saturation forecast: Around June 2028. 9 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.534.16#54
2Claude Opus 4.731.14#85
3DeepSeek V4 Pro29.77#196
4GLM-5.129.56#198
5Claude Sonnet 4.628.89#173
6DeepSeek V4 Flash28.39#267
7Qwen 3.6 Plus27#238
8MiniMax-M2.724.69#539
9Kimi K2.616.39#192

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=deepweb-bench-calibration · How It Works · Data refreshed daily, snapshot 2026-11-21.