EXAONE 1.0 (LG) and Inflection-2 (Inflection AI) are both frontier AI models. EXAONE 1.0 was published in December 2021 and Inflection-2 in November 2023.
Inflection-2 was trained on 10^25 FLOP, about 5.9x the compute of EXAONE 1.0 at 1.7 x 10^24 FLOP. Training compute is the closest available proxy for how much was invested in a model, though it says nothing on its own about how well that compute was spent.
These two models share no benchmark on which both have been scored, so no direct performance comparison is possible here. The specification table below is a comparison of inputs, not of results.
| Field | EXAONE 1.0 | Inflection-2 |
|---|---|---|
| Organization | LG | Inflection AI |
| Published | Dec 14, 2021 | Nov 22, 2023 |
| Training compute | 1.7 x 10^24 FLOP | 10^25 FLOP |
| Parameters | 300B | — |
| Training hardware | — | NVIDIA H100 SXM5 80GB |
| Chips used | — | 5,000 |
| Training cost (2023 USD) | $3M | $13M |
| Training power draw | — | 6.9 MW |
| Accessibility | Unreleased | Hosted access (no API) |
| Open weights | No | No |
| Country | South Korea | United States |
