Open models, in actual practice, don't match up to even one or two generation prior models from Anthropic/OpenAI/Google. They've clearly been trained on the benchmarks. Entirely possible it was by mistake, but it's definitely happening.

Which open model has the same performance as Opus 4.7?

I wouldn't say the same but it's pretty close. At this point I'm convinced that they'll continue running the marketing machine and people due to FOMO will keep hopping onto whatever model anthropic releases.

I think that's precisely why they're paying thousands of people in those other jobs to perform their tasks while collecting new data. Software was easiest because its already mostly written down, but other jobs can be quantized with enough data points. Just give it time