BALSAM - Factuality: leaderboard
Metric: Overall score (0-100, LLM-judged generation and multiple choice). Source: benchmarks.ksaa.gov.sa. 29 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemma 4 31B (IT) | 29.11 |
| 2 | Command A 03 2025 | 26.56 |
| 3 | Qwen 3 235B A22B | 25.67 |
| 4 | GLM-5 (Thinking) | 25.44 |
| 5 | GLM-4.7 (Reasoning) | 25.44 |
| 6 | Jais-2-70B-Chat | 25.44 |
| 7 | DeepSeek V3.2 | 24.78 |
| 8 | GPT-5.2 | 24.55 |
| 9 | Claude Opus 4.5 | 23.11 |
| 10 | Fanar-C-2-27B | 22.11 |
| 11 | Command-R+ | 21 |
| 12 | Llama 4 Maverick Instruct | 20.89 |
| 13 | Kimi K2 Instruct (0905) | 20.89 |
| 14 | Yehia-7B-preview | 20.33 |
| 15 | AceGPT-v2-8B-Chat | 20.22 |
Interactive version: theaggregate.ai/benchmark?slug=balsam-factuality · How It Works · Data refreshed daily, snapshot 2026-09-19.