BALROG TextWorld (LLM) — leaderboard
Metric: Progress (%). Source: balrogai.com. 34 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) (Thinking) | 75.7 |
| 2 | Gemini 3.1 Pro (Preview) | 66.5 |
| 3 | Grok 4 | 62.9 |
| 4 | Gemini 3 Pro | 60.2 |
| 5 | Claude Opus 4.5 (Thinking) | 59 |
| 6 | Claude Opus 4.5 | 51.4 |
| 7 | Gemini 3 Flash | 50.2 |
| 8 | Gemini 2.5 Pro (03-25) | 49.2 |
| 9 | Claude Haiku 4.5 | 49 |
| 10 | Claude 3.5 Sonnet (20241022) | 42.1 |
| 11 | GPT-4o (2024-05-13) | 39.3 |
| 12 | Grok 3 | 32.5 |
| 13 | GPT-5 (Minimal) | 30.6 |
| 14 | Gemini 2.5 Flash | 29.6 |
| 15 | DeepSeek R1 | 21.8 |
Interactive version: theaggregate.ai/benchmark?slug=balrog-textworld-llm · How the rankings work · Data refreshed daily, snapshot 2026-07-22.