MLCR-AA Conciseness
Professional Work
Share of MLCR-AA responses meeting the length and answer-presence requirements. This does not measure answer correctness and is excluded from scalar capability fitting.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Sol | 100.0 |
| Nova Lite | 100.0 |
| Command A+ (May 2026, BF16) | 99.4 |
| Gemini 3.1 Pro Preview | 99.4 |
| GPT-5.4 Mini | 99.4 |
| Grok 4.6 | 99.4 |
| Nemotron 3 Super 120B A12B | 99.4 |
| Gemini 3.1 Flash-Lite Preview | 98.9 |
| Gemini 3 Flash Preview | 98.9 |
| Grok 4.3 | 98.9 |
| Qwen3.7-Plus | 98.9 |
| Gemini 2.5 Flash-Lite | 98.3 |
| Gemini 3.5 Flash-Lite | 98.3 |
| Gemini 3.6 Flash | 98.3 |
| Gemini 3.7 Flash | 98.3 |
| Grok 4.5 | 98.3 |
| Llama 4 Maverick Instruct | 98.3 |
| Opus 4.7 | 97.8 |
| Gemini 2.5 Flash | 97.8 |
| Gemini 3.5 Flash | 97.8 |
| GPT-4o Mini | 97.8 |
| GPT-5.6 Sol | 97.8 |
| GPT-6 Luna | 97.8 |
| Grok 4.7 | 97.8 |
| Kimi K2.7 Code | 97.8 |
| Nemotron 3 Ultra 550B A55B | 97.8 |
| DeepSeek-V4-Flash-0731 | 97.2 |
| DeepSeek-V4-Flash | 97.2 |
| Gemini 3.8 Flash | 97.2 |
| Nemotron 3.5 Lightning 30B A3B (precision unspecified) | 97.2 |
| DeepSeek V4.1 Flash | 96.7 |
| GLM-5.3 Flash | 96.7 |
| GPT-5.4 Nano | 96.7 |
| GPT-5.5 | 96.7 |
| GPT-5.6 Luna | 96.7 |
| GPT-5.6 Terra | 96.7 |
| GPT-6 Astra | 96.7 |
| Inkling | 96.7 |
| Step 5 Preview | 96.7 |
| Opus 4.8 | 96.1 |