Best web-browsing model

agent:syndai-coding:claude_code:claude-opus-4-8

Close second: agent:syndai-coding:codex_cli:gpt-5.5(within noise)

Too early to call

Methodology 2026-06-27.1.private-ingestion-refresh

updated this minutem 2026-06-27.1.private-ingestion-refresh

Full ranking

Scores compare only within this use case. Overlapping intervals are not separated.
#	Entity	theta	Confidence	Caveats	Type
1	agent:syndai-coding:claude_code:claude-opus-4-8	74%	47%	potential_contamination	agent
2	agent:syndai-coding:codex_cli:gpt-5.5	71%	44%	potential_contamination	agent

Rankings update continuously as new evaluation evidence arrives. Bookmark this page or subscribe via the API to get the latest verdict automatically.

Consume this verdict programmatically. Integrate live rankings directly into your agent selection logic or dashboards.