APEX (Mercor)
Professional Work · 2025-10-01
Mercor's APEX evaluates models on realistic professional tasks authored by veteran practitioners across four occupations -- investment banking associate, management consultant, big law associate, and primary care physician -- with 100 tasks each. Every task is answered eight times and graded by a judge model, and the reported score is the mean.
Top models (higher is better)
| Model | Score |
|---|---|
| Gemini 4 Argon | 32.4 |
| Claude Fable 5.1 | 26.7 |
| Claude Opus 5.5 | 23.9 |
| Muse Spark 1.2 | 23.7 |
| Claude Opus 5 | 22.4 |
| GPT-5.6 Sol Pro | 22.1 |
| Muse Wonder | 22.1 |
| Claude Fable 5 | 21.5 |
| Muse Spark 1.1 | 21.2 |
| Muse Spark 1.3 | 21.2 |
| Muse Spark Open | 20.9 |
| GPT-6.1 Sol | 20.1 |
| Gemini 3.8 Flash | 19.9 |
| GPT-6 Astra | 19.9 |
| GPT-5.6 Terra | 17.7 |
| GPT-6.1 Terra | 17.7 |
| Gemini 3.7 Flash | 17.6 |
| Grok 4.6 | 17.6 |
| Claude Sonnet 5.5 | 17.5 |
| DeepSeek V4.1 Flash 1017 | 17.4 |
| GLM-5.3 | 16.6 |
| GPT-5.6 Luna | 16.6 |
| Kimi K3 | 16.6 |
| Grok 4.7 | 16.3 |
| GPT-5.5 | 16.2 |
| Opus 4.8 | 16.1 |
| DeepSeek V4.1 Flash | 15.9 |
| DeepSeek V4 Pro 0813 | 15.9 |
| Claude Haiku 5.5 | 15.2 |
| Gemini 3.5 Flash | 15.2 |
| GPT-5.3-Codex | 15.2 |
| Opus 4.6 | 14.9 |
| Gemini 3.6 Flash | 14.8 |
| GPT-5.4 | 14.6 |
| GPT-6 Sol | 14.5 |
| MiniMax M3.1 Flash | 14.4 |
| Opus 4.7 | 13.7 |
| GPT-6 Luna | 13.4 |
| Inkling | 13.3 |
| Beam 1.0 501B A23B | 13.0 |