4 ms·
DeepSWE Benchmark Results for GPT 5.6
- tomerlir 3mo agoborderline better
- theanonymousone 3mo agoI wouldn't call it borderline. If we take the chart seriously (if), it means for every budget, the best performance is achieved by a 5.6 variant.
loading thread…