Research group developing AI tools for understanding climate communication at scale.
We build and release models that detect and classify 100+ types of climate contrarianism claims in text, based on the peer-reviewed CARDS taxonomy:
Coan et al. (2026). Large language model reveals an increase in climate contrarian speech in the United States Congress. Communications Sustainability. doi:10.1038/s44458-025-00029-z
| Model | Parameters | Link |
|---|---|---|
| CARDS-Qwen3.5-4B | 4B | C3DS/CARDS-Qwen3.5-4B |
| CARDS-Qwen3.5-9B | 9B | C3DS/CARDS-Qwen3.5-9B |
| CARDS-Qwen3.5-27B | 27B | C3DS/CARDS-Qwen3.5-27B |
| Model | Q4_K_M | Q8_0 | Link |
|---|---|---|---|
| CARDS-Qwen3.5-4B | 2.7 GB | 4.2 GB | C3DS/CARDS-Qwen3.5-4B-GGUF |
| CARDS-Qwen3.5-9B | 5.6 GB | 9.8 GB | C3DS/CARDS-Qwen3.5-9B-GGUF |
| CARDS-Qwen3.5-27B | 16 GB | 28.6 GB | C3DS/CARDS-Qwen3.5-27B-GGUF |
ollama run exec3ds/cards-qwen3.5 # 4B default
ollama run exec3ds/cards-qwen3.5:9b
ollama run exec3ds/cards-qwen3.5:27b
Just give it text — the system prompt and chain-of-thought reasoning are built in:
ollama run exec3ds/cards-qwen3.5 "CO2 is just a trace gas. The sun drives climate."