BALSAM - Creative Writing: 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) | 68.79 |
| 2 | GLM-5 (Thinking) | 54.57 |
| 3 | GLM-4.7 (Reasoning) | 52.76 |
| 4 | GPT-5.2 | 52.27 |
| 5 | Kimi K2 Instruct (0905) | 51.19 |
| 6 | Command A 03 2025 | 51.1 |
| 7 | DeepSeek V3.2 | 50.75 |
| 8 | Fanar-C-2-27B | 50.68 |
| 9 | MiniMax M3 (Reasoning) | 50.55 |
| 10 | AceGPT-v2-8B-Chat | 50.35 |
| 11 | Grok 4.1 Fast (Reasoning) | 49.89 |
| 12 | Llama 3.3 70B Instruct | 49.79 |
| 13 | Qwen 3 235B A22B | 49.54 |
| 14 | Llama 4 Maverick Instruct | 48.78 |
| 15 | Claude Opus 4.5 | 48.77 |
Interactive version: theaggregate.ai/benchmark?slug=balsam-creative-writing · How It Works · Data refreshed daily, snapshot 2026-09-19.