Qwen-AgentWorld: leaderboard
A world model predicts environment dynamics based on current observations and actions, serving as a core cognitive mechanism for reasoning and planning. In this work, we investigate how world modeling based on language models can further push the boundaries of general agents. (i) We first focus on building foundation models for agentic environment simulation. We introduce Qwen-AgentWorld-35B-A3B a
Metric: Overall (imported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 58.25 | #79 |
| 2 | Claude Opus 4.6 | 57.8 | #63 |
| 3 | Claude Opus 4.8 | 56.59 | #37 |
| 4 | Claude Sonnet 4.6 | 56.04 | #90 |
| 5 | Qwen 3.5 397B A17B | 54.74 | #146 |
| 6 | Gemini 3.1 Pro (Preview) | 54.57 | #58 |
| 7 | Kimi K2.6 | 53.42 | #104 |
| 8 | DeepSeek V4 Pro | 52.97 | #102 |
| 9 | Qwen 3.6 Max Preview | 52.42 | #85 |
| 10 | GLM-5.1 | 51.31 | #97 |
| 11 | Qwen 3.6 Plus | 50.81 | #120 |
| 12 | Qwen 3.5 35B A3B | 47.73 | #256 |
| 13 | MiniMax-M2.7 | 46.12 | #258 |
| 14 | Qwen 3.6 35B A3B | 42.88 | #226 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=qwen-agentworld · How It Works · Data refreshed daily, snapshot 2026-10-20.