Same standard of work. About 6.4× the energy.
Grok 4.7 scores higher than Llama 3-70B on Epoch's Capabilities Index, 154.37 against 122.94, across 58 benchmarks covering coding, agents, long context and writing, while using about 6.4× less energy per answer. Grok 4.7 has not faced human head-to-head voting, where Llama 3-70B carries 156,876 votes, worth weighing.
| Grok 4.7 | Llama 3-70B | |
|---|---|---|
| Energy per answer | 0.285 Wh | 1.84 Wh |
| Carbon released | 0.13 g CO₂ | 0.84 g CO₂ |
| Same carbon as driving | under a metre | 3 metres (11 feet) |
| Energy rank | 21 of 277 | 148 of 277 |
| Capability (58 benchmarks) | 154.4 | 122.9 |
| Human vote rating | - | 1221 |
| Votes cast | - | 156,876 |
| Confidence | Calculated | Calculated |
| Released | 2026-09-21 | 2024-04-18 |
Energy is calculated the same way for every model on this site, from public data and calibrated against laboratory measurement, accurate to roughly a factor of two to three, which is why we only ever state a difference of 5× or more. Capability is Epoch AI's Capabilities Index, one score combining 58 separate benchmarks; the vote rating comes from blind human comparisons. The full method is here.