SMH-Bench (Interactive Agent): leaderboard
SMH-Bench (2026) tests LLM smart-home agents on 1,100 human-audited tasks in HomeEnv, an executable simulator of homes from small apartments to a 31-room, 135-device house; success means the final device state meets the goal without touching unrelated devices, with GPT-5 judging clarifications and query answers. In the Environment-Interactive Agent setting the model starts with partial observability and must query and control the home through tools in a ReAct loop. The board is the success rate in percent over all tasks; higher is better.
Metric: Success rate (%), instance-weighted over 1,100 tasks in 7 capability categories; Environment-Interactive Agent: a ReAct loop that starts from partial observability and queries and controls the HomeEnv simulator through tools under a call budget; final home state checked against the goal with untouched devices preserved; clarification and query answers judged by GPT-5; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 85.2 | #94 |
| 2 | Claude Sonnet 4.6 | 84.1 | #161 |
| 3 | GPT-5.4 | 79.6 | #123 |
| 4 | Qwen 3.5 397B A17B | 77.8 | #246 |
| 5 | Claude Haiku 4.5 | 76.2 | #568 |
| 6 | DeepSeek V3.2 (Thinking) | 75.5 | #406 (DeepSeek V3.2) |
| 7 | GPT-5.4 Mini | 75.3 | #389 |
| 8 | GLM-5 | 70.6 | #260 |
| 9 | Qwen 3.5 Plus | 70 | #232 |
| 10 | Qwen 3.5 9B | 64.3 | #803 |
| 11 | Qwen 3.5 4B | 56.5 | #1175 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=smh-bench-interactive-agent · How It Works · Data refreshed daily, snapshot 2026-11-13.