LLM Stats (Agents' Last Exam): leaderboard
Metric: Score (%). Source: llm-stats.com. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-6 | 59.3 |
| 2 | GPT-5.6 Sol | 52.7 |
| 3 | Qwen 3.8 Max | 52.4 |
| 4 | Qwen3.8-Flash-Next | 51.2 |
| 5 | Qwen 3.8 Flash | 51.2 |
| 6 | GPT-5.6 Terra | 50.4 |
| 7 | GPT-5.6 Luna | 50.3 |
| 8 | Qwen 3.8 27B | 42.9 |
| 9 | Seed 2.1 Pro | 41.4 |
| 10 | DeepSeek V4.1 Flash | 31.8 |
| 11 | GLM-5.3 | 28.5 |
| 12 | DeepSeek-V4-Flash-Vision-Exp | 27.3 |
| 13 | GLM-5.3 Flash | 26.3 |
| 14 | Gemini 3.7 Flash | 26.3 |
| 15 | DeepSeek V4 Pro (0813) | 25.7 |
Interactive version: theaggregate.ai/benchmark?slug=llm-stats-agents-last-exam · How It Works · Data refreshed daily, snapshot 2026-09-19.