鹈鹕骑车标本馆PELICAN MAP · AN AI FIELD GUIDE
首页 / 馆藏 / 标本详情
社区记录 · 2025-06-06

34 pelicans; Gemini 2.5 Pro preview won

GPT-4.1 mini 当裁判打了 560 场，花费约 18 美分。

34 pelicans; Gemini 2.5 Pro preview won · ASCII art Leaderboard table showing AI model rankings with columns for Rank, Model, Elo, Matches, Wins, and Win Rate. Top models include: 1. gemini-2.5-pro-preview-05-06 (1800.4 Elo, 100.0% win rate), 2. gemini-2.5-pro-preview-03-25 (1769.9 Elo, 97.0% win rate), 3. o3 (1767.8 Elo, 90.9% win rate), 4. claude-4-sonnet (1737.9 Elo, 90.9% win rate), continuing down to 34. llama-3.3-70b-instruct (1196.2 Elo, 0.0% win rate). Footer shows "Total models: 34, Total matches: 560".
ASCII art Leaderboard table showing AI model rankings with columns for Rank, Model, Elo, Matches, Wins, and Win Rate. Top models include: 1. gemini-2.5-pro-preview-05-06 (1800.4 Elo, 100.0% win rate), 2. gemini-2.5-pro-preview-03-25 (1769.9 Elo, 97.0% win rate), 3. o3 (1767.8 Elo, 90.9% win rate), 4. claude-4-sonnet (1737.9 Elo, 90.9% win rate), continuing down to 34. llama-3.3-70b-instruct (1196.2 Elo, 0.0% win rate). Footer shows "Total models: 34, Total matches: 560". 附件原始来源 ↗

原目录标签：L1。这是整理标签，不能作为统一条件下的模型能力评分。

Markdown 资料

继续观察

查看全部馆藏 ↗