Ranking Updates
Every snapshot, in reverse order. Models tested, categories covered, and what was new.
- OpenAI GPT-6.1
- Anthropic Claude Opus 5.5
- Google Gemini 3.1
New models this month: GPT-6.1 (from GPT-5.6) and Claude Opus 5.5 (from Opus 5); Gemini 3.1 is unchanged. Opus 5.5 names about 15–20 products per answer, down from 25–40, so long-tail tools that only Claude listed have dropped off many boards, and leaderboards are shorter this month.
- OpenAI GPT-5.6
- Anthropic Claude Opus 5
- Google Gemini 3.1
No model changes between this snapshot and August 2026. 33 categories and 91 sub-categories, 10 runs per model per category.
- OpenAI GPT-5.6
- Anthropic Claude Opus 5
- Google Gemini 3.1
Anthropic updated from Claude Opus 4.8 to Claude Opus 5 and OpenAI updated from GPT-5.5 to GPT-5.6 between this snapshot and July 2026. Sub-categories expanded from 23 to 91. 33 categories, 10 runs per model per category.
- OpenAI GPT-5.5
- Anthropic Claude Opus 4.8
- Google Gemini 3.1
Leaderboards now rank by prominence — a 1–100 score combining visibility and list position via a modified Dowdall count. Average position breaks ties. Visibility and average position remain available on hover. Previous months were re-ordered based on prominence for fair comparisons.
- OpenAI GPT-5.5
- Anthropic Claude Opus 4.8
- Google Gemini 3.1
Anthropic updated from Claude Opus 4.7 to Claude Opus 4.8 between this snapshot and May 2026. 33 categories, 10 runs per model per category.
- OpenAI GPT-5.5
- Anthropic Claude Opus 4.7
- Google Gemini 3.1
Model updates since April 2026: OpenAI (GPT-5.4 → GPT-5.5); Anthropic (Claude Opus 4.6 → Claude Opus 4.7); Google (Gemini 3 → Gemini 3.1). 33 categories, 10 runs per model per category.
- OpenAI GPT-5.4
- Anthropic Claude Opus 4.6
- Google Gemini 3
First three-model snapshot. Expanded from 14 to 33 categories. 10 runs per model per category.
- OpenAI GPT-5.2 (current)
- OpenAI GPT-4 (comparison baseline)
Expanded from 14 to 33 categories. Rankings compared GPT-5.2 recommendations against a GPT-4 baseline. 10 runs per model per category.
- OpenAI GPT-5.2 (current)
- OpenAI GPT-4 (comparison baseline)
Initial launch. 14 categories. Rankings compared GPT-5.2 recommendations against a GPT-4 baseline. 10 runs per model per category.