Updated daily at 05:00 UTC
Ranking
Best AI models for Professional knowledge work
Producing real deliverables (documents, spreadsheets, analyses) of the kind professionals are paid for, judged against expert work.
Ranked models
Based on 1 benchmark
| Rank | Model | Company | Score | Coverage | Released |
|---|---|---|---|---|---|
| 1 | DeepSeek-V3.2-Exp | DeepSeek | 42.9 | 1 of 1 | 29 Sep 2025 |
| 2 | Gemini 2.5 Flash | Google DeepMind | 41.1 | 1 of 1 | 17 Apr 2025 |
| 3 | Claude Sonnet 4 Retired | Anthropic | 33.1 | 1 of 1 | 22 May 2025 |
| 4 | Claude 3.7 Sonnet Retired | Anthropic | 30.9 | 1 of 1 | 24 Feb 2025 |
| 5 | Gemini 2.5 Pro | Google DeepMind | 30.3 | 1 of 1 | 25 Mar 2025 |
| 6 | Claude 3.5 Sonnet Retired | Anthropic | 24.0 | 1 of 1 | 20 Jun 2024 |
| 7 | Gemini 2.0 Flash Retired | Google DeepMind | 11.4 | 1 of 1 | 11 Dec 2024 |
| 8 | GPT-4o | OpenAI | 8.6 | 1 of 1 | 13 May 2024 |
| 9 | Llama 3.1 405B | Meta | 7.4 | 1 of 1 | 23 Jul 2024 |
| 10 | Llama 3.1 70B | Meta | 6.9 | 1 of 1 | 23 Jul 2024 |
| 11 | Qwen2.5-72B-Instruct | Alibaba Group | 5.7 | 1 of 1 | 19 Sep 2024 |
| 12 | Gemini 1.5 Pro Retired | Google DeepMind | 3.4 | 1 of 1 | 15 Feb 2024 |
| 13 | Amazon Nova Pro | Amazon | 1.7 | 1 of 1 | 3 Dec 2024 |
| 14 | Qwen2-72B-Instruct | Alibaba Group | 1.1 | 1 of 1 | 7 Jun 2024 |
