xAI has completed the full API rollout of Grok 4.3 — and the changes are significant. Input prices drop by roughly 40 percent, the context window expands to one million tokens, and for the first time, Grok can natively process video.
The numbers
Grok 4.3 now costs $1.25 per million input tokens and $2.50 per million output tokens (up to 200,000 input tokens, after which costs double). For comparison: Claude Opus 4.7 sits at $15 input and $75 output. Grok is orders of magnitude cheaper — though with notably lower benchmark scores.
On the Artificial Analysis Intelligence Index, Grok 4.3 scores 53. That’s solid for its price tier, but far from frontier models. xAI clearly isn’t going for “best model in the world” — they’re going for “best price-performance ratio for agents.”
Video input as a differentiator
New to the lineup is the ability to process videos directly — a feature neither Claude nor GPT-5.5 offer in this form. There’s also the ability to generate presentation slides directly in chat. xAI is increasingly building Grok as a multimodal all-rounder that might not hit the highest benchmarks but aims to be practical for everyday use.
What the price war means
xAI’s aggressive pricing ramps up pressure on every provider. As agent-based workflows become the norm, the cost per token becomes critical — because agents consume far more tokens than individual chat queries. xAI is positioning itself as the affordable workhorse for companies that want their agents running around the clock.
For Claude users, nothing changes in the short term. But long-term, Grok 4.3 shows where things are headed: prices are falling, context windows are growing, and competition will be decided by everyday utility — not benchmarks.
Sources: VentureBeat, Artificial Analysis, Apiyi