AI Energy Score (Text Generation): leaderboard
Hugging Face energy efficiency ratings for text generation models across GPU tiers, benchmarked on standardized datasets to compare power consumption.
Metric: Energy Score (1-5). Source: huggingface.co. 62 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 5 | #454 |
| 2 | Qwen 3 30B A3B | 5 | #493 |
| 3 | Granite 4.0 Micro | 5 | #882 |
| 4 | DeepSeek R1 Distill Qwen 32B | 5 | #655 |
| 5 | Qwen 3 0.6B | 5 | #1459 |
| 6 | GLM 4.5 Air | 5 | #392 |
| 7 | DeepSeek R1 Distill Llama 70B | 5 | #594 |
| 8 | gemma-3-1B-pt Base | 5 | #1540 |
| 9 | Gemma 3 270M | 5 | #1571 |
| 10 | Llama 3 70B | 5 | #631 |
| 11 | Phi-4-reasoning-plus (Thinking) | 5 | |
| 12 | SmolLM3-3B | 5 | #1252 |
| 13 | Llama 2 70B Base | 5 | #1025 |
| 14 | QwQ 32B-Preview | 5 | #651 |
| 15 | Phi-1.5 | 5 | #1549 |
Interactive version: theaggregate.ai/benchmark?slug=ai-energy-score-text-generation · How It Works · Data refreshed daily, snapshot 2026-10-20.