Grok 4.5 Ranks Fourth On Intelligence Index At Lower Cost
Benchmark trackers say SpaceXAI's new model climbed near the front of two leaderboards while undercutting pricier rivals.
Artificial Analysis says SpaceXAI's Grok 4.5 scored 54 on its Intelligence Index, good for fourth place behind models from Anthropic and OpenAI, while Voxelbench says the model also landed fourth on its leaderboard. The same Artificial Analysis writeup says Grok 4.5 improved 16 points over Grok 4.3 and matched GPT-5.5 in its Coding Agent Index within the Grok Build harness at lower cost. On X, that mix of higher rankings and cheaper pricing quickly turned into a victory lap for Elon Musk and xAI.
Turns out that the cursor acquisition was insanely cheap
Commentary on X
Highest rankedCombined views
46.8K
2 posts, first seen 3h ago