Loading blueprint versions...
Please wait while we gather all the unique runs for this blueprint.
Please wait while we gather all the unique runs for this blueprint.
Please wait while we prepare the detailed comparison.
DTEF: Predict response distributions for Pakistan. Source: Global Dialogues GD1 (https://github.com/collect-intel/global-dialogues)
Average key point coverage extent for each model across all prompts.
| Prompts vs. Models | Claude 3.7 Sonnet | Claude 3.5 Haiku | Claude Haiku 4.5 | Claude Sonnet 4 | Gemini 2.5 Flash | Gemma 3 12b It | Llama 3 70b Instruct | Llama 4 Maverick | Mistral Large 2411 | Mistral Medium 3 | Mistral Nemo | GPT 4.1 | GPT 4.1 Mini | GPT 4.1 Nano | GPT 4o | GPT 4o Mini | GPT OSS 120b | GPT OSS 20b | GLM 4.5 | Qwen3 30b A3B Instruct 2507 | Qwen3 32b | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Score | 10th 81.8% | 3rd 84.8% | 4th 83.8% | 5th 83.3% | 21st 74.5% | 13th 79.3% | 2nd 86.3% | 19th 75.8% | 17th 77.3% | 13th 79.3% | 1st 87.0% | 8th 82.5% | 5th 83.3% | 18th 76.8% | 9th 82.3% | 11th 81.3% | 20th 74.8% | 16th 78.3% | 12th 81.0% | 7th 82.7% | 15th 79.0% | |
| 62.3% | 59% | 65% | 63% | 54% | 67% | 54% | 64% | 51% | 62% | 58% | 72% | 58% | 68% | 72% | 63% | 71% | 55% | 60% | 68% | 65% | 59% | |
| 89.2% | 89% | 97% | 86% | 93% | 83% | 87% | 96% | 90% | 91% | 89% | 97% | 97% | 90% | 72% | 97% | 84% | 75% | 83% | 88% | 94% | 96% | |
| 84.2% | 86% | 86% | 90% | 89% | 84% | 84% | 90% | 78% | 82% | 86% | 90% | 84% | 84% | 67% | 83% | 90% | 83% | 80% | 84% | 88% | 81% | |
| 87.0% | 93% | 91% | 96% | 97% | 64% | 92% | 95% | 84% | 74% | 84% | 89% | 91% | 91% | 96% | 86% | 80% | 86% | 90% | 84% | 84% | 80% |