Same standard of work. About 8.4× the energy.
GPT-5.6 Luna scores higher than MiniMax-M3 on Epoch's Capabilities Index, 156.5 against 146.87, across 58 benchmarks covering coding, agents, long context and writing, while using about 8.4× less energy per answer. In blind head-to-head voting, people rate them 1431 and 1431 respectively.
| GPT-5.6 Luna | MiniMax-M3 | |
|---|---|---|
| Energy per answer | 2.61 Wh | 22.0 Wh |
| Carbon released | 1.2 g CO₂ | 10 g CO₂ |
| Same carbon as driving | 5 metres (16 feet) | 41 metres (133 feet) |
| Energy rank | 169 of 278 | 250 of 278 |
| Capability (58 benchmarks) | 156.5 | 146.9 |
| Human vote rating | 1431 | 1431 |
| Votes cast | 41,317 | 68,611 |
| Confidence | Day-one estimate | Calculated |
| Released | 2026-07-09 | 2026-06-01 |
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.