BALSAM - Reading Comprehension: 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) | 68.41 |
| 2 | DeepSeek V3.2 | 63.43 |
| 3 | Command A 03 2025 | 59.8 |
| 4 | GPT-5.2 | 59.13 |
| 5 | DeepSeek V4 Pro (Reasoning) | 58.55 |
| 6 | Command-R+ | 58.16 |
| 7 | Qwen 3 235B A22B | 57.43 |
| 8 | AceGPT-v2-8B-Chat | 55.97 |
| 9 | Yehia-7B-preview | 54.72 |
| 10 | Llama 4 Maverick Instruct | 52.58 |
| 11 | GLM-5 (Thinking) | 52.5 |
| 12 | Fanar-C-2-27B | 52.34 |
| 13 | Claude Opus 4.5 | 52.06 |
| 14 | Kimi K2 Instruct (0905) | 51.11 |
| 15 | GLM-4.7 (Reasoning) | 50.97 |
Interactive version: theaggregate.ai/benchmark?slug=balsam-reading-comprehension · How It Works · Data refreshed daily, snapshot 2026-09-19.