APEX-Agents (DualViewEval Study): leaderboard
The DualViewEval paper's own end-to-end re-measurement of 26 APEX-Agents model configurations, run in full to validate its estimator rather than estimated from a subset. Higher is better. Source: arxiv.org/abs/2609.18909.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Around February 2027. 26 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 52.27 | #40 |
| 2 | Gemini 3.5 Flash | 45.19 | #67 |
| 3 | GPT-5.4 | 44.83 | #84 |
| 4 | GPT-5.5 | 43.68 | #33 |
| 5 | GLM-5.2 | 43.25 | #79 |
| 6 | Kimi K2.6 | 37.62 | #113 |
| 7 | Claude Opus 4.7 | 35.04 | #52 |
| 8 | GLM-5.1 | 34.62 | #109 |
| 9 | Qwen 3.7 Max | 32.45 | #86 |
| 10 | Seed 2.1 Pro | 32.01 | #75 |
| 11 | Gemini 3 Flash | 31.81 | #105 |
| 12 | Claude Opus 4.6 | 30.83 | #68 |
| 13 | Seed 2.0 Pro | 29.29 | #123 |
| 14 | Qwen 3.7 Plus | 24.25 | #99 |
| 15 | Kimi K2.5 | 23.53 | #153 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=apex-agents-dualvieweval-study · How It Works · Data refreshed daily, snapshot 2026-11-06.