Taiwan Exams - TVE Joint Entrance Exam (ROC 115): leaderboard

Metric: Points Scored (%). Source: huggingface.co. 20 models tracked.

Top models

#ModelScore
1Codex CLI (gpt-6-astra, medium)98.41
2Claude Code (claude-opus-5, xhigh)98.21
3Codex CLI (gpt-6-astra, high)98.15
4Claude Code (claude-opus-5, high)98.08
5Codex CLI (gpt-6-astra, xhigh)97.96
6Codex CLI (gpt-6-astra, low)97.86
7Codex CLI (gpt-5.6-sol, xhigh)97.66
8Claude Code (claude-opus-5, medium)97.61
9Claude Code (claude-opus-5, low)97.35
10Codex CLI (gpt-5.6-sol, high)97.15

Interactive version: theaggregate.ai/benchmark?slug=taiwan-exams-tve-joint-entrance-exam-roc-115 · How It Works · Data refreshed daily, snapshot 2026-09-19.