# Agent's Last Exam (pass@1) — AI model rankings

A hard set of desktop and operating-system tasks an AI agent has to finish by looking at the screen and working the machine itself. The score is the share it passes outright — partial credit does not count. Higher is better.

15 tracked models have a published Agent's Last Exam (pass@1) score. Higher is better. Scores come from published lab reports and benchmark sources; recorded sources appear beside the scores.

## Ranking

| Rank | Model | Developer | Score | Source | Released |
| --- | --- | --- | --- | --- | --- |
| 1 | GPT-6.1 Terra | OpenAI | 60% | Lab | Oct 26 2026 |
| 2 | GPT-6 Astra | OpenAI | 59.3% | Lab | Sep 3 2026 |
| 3 | GPT-6 Sol | OpenAI | 56.4% | Lab | Sep 22 2026 |
| 4 | GPT-5.6 Sol | OpenAI | 52.7% | Lab | Jun 26 2026 |
| 5 | GPT-5.6 Terra | OpenAI | 50.4% | Lab | Jun 26 2026 |
| 6 | GPT-5.6 Luna | OpenAI | 50.3% | Lab | Jun 26 2026 |
| 7 | Gemini 4 Argon | Google | 39.5% | Lab | Sep 30 2026 |
| 8 | DeepSeek-V4.1-Flash-1017 | DeepSeek | 35.4% | Lab | Oct 17 2026 |
| 9 | DeepSeek-V4.1-Flash | DeepSeek | 31.8% | Lab | Sep 10 2026 |
| 10 | GLM-5.3 | Z.ai | 28.5% | Lab | Aug 14 2026 |
| 11 | Gemini 3.7 Flash | Google | 26.3% | Lab | Aug 13 2026 |
| 11 | GLM-5.3-Flash | Z.ai | 26.3% | Lab | Aug 26 2026 |
| 13 | Qwen3.8-Flash-Next | Qwen | 24.3% | Lab | Aug 26 2026 |
| 14 | Gemini 3.6 Flash | Google | 24.2% | Lab | Jul 21 2026 |
| 15 | Qwen3.8-27B | Qwen | 20.4% | Lab | Aug 14 2026 |


---

Canonical page: https://aireleasetracker.com/benchmark/agents-last-exam
Site index: https://aireleasetracker.com/llms.txt
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.