Grok 4 still ranks #1 on the FinSearchComp Benchmark
It's a first expert-level benchmark for financial search & reasoning
Grok 4 is very close to approaching the human experts' level....
FSD V13.2.2 ➡️ V13.2.7
Drive 1, 25 miles. City & highway mix.
FIRST IMPRESSIONS:
At this point, there really isn't much to say. I pressed a button on a touch screen from the parking lot at my apartment complex, sat on my ass, pressed the door open button to get out after the car parallel parked me at destination. I could’ve napped if I was allowed to.
Of course there are still certain edge cases we know it hasn’t solved yet, but at this point, I go entire months without an intervention beyond disengaging to park. Seriously.
@Tesla_AI I can’t believe you guys built this 😂😂. It’s too much man, too much.
Grateful that I’ve gotten to watch this happen over the last 14 months. Can’t imagine what this feels like for the real OGs.