TerraLogic - Instruction Following: leaderboard
Metric: Instruction-following accuracy (InstAcc) (%) in the step-by-step setting over the 545 TerraLogic optical, SAR and infrared remote-sensing tasks, each model driving the HieraPlan tool agent under a ReAct-style protocol in OpenCompass; higher is better. Source: arxiv.org. Saturation forecast: Around June 2028. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4o | 79.4 |
| 2 | Qwen 2.5 7B Instruct | 77.19 |
| 3 | Llama 3 70B Instruct | 67.89 |
| 4 | DeepSeek V3 | 65.84 |
| 5 | Yi-1.5-6B-Chat | 63.38 |
| 6 | Qwen 2.5 32B Instruct | 62.83 |
| 7 | InternLM3-8B-Instruct | 61.26 |
| 8 | Mistral 7B Instruct (v0.2) | 60.03 |
| 9 | Gemini 2.5 Flash | 59.09 |
| 10 | GPT-3.5 | 59.05 |
| 11 | Phi-3 Mini 4K Instruct | 55.45 |
Interactive version: theaggregate.ai/benchmark?slug=terralogic-instruction-following · How It Works · Data refreshed daily, snapshot 2026-09-29.