Beckmann

Dossier · Ongoing

Grok 4.x: SpaceXAI's price-performance race

From Grok 4.5 in July to Grok 4.6 in August 2026: how SpaceXAI positions its coding models on price and performance against the competition.

SpaceXAI has driven a price war among AI coding models with its Grok 4 line since July 2026: Grok 4.5 launched by sharply undercutting Anthropic’s pricing but was initially blocked in Europe over its classification as a systemic-risk model, until access was unlocked in Germany in mid-July.

With Grok 4.6 in August 2026, SpaceXAI holds pricing steady at 2 and 6 dollars per million tokens while extending post-training specifically for agentic tasks, tying OpenAI’s GPT-5.6 Sol according to independent analysis firm Artificial Analysis.

Timeline

  1. Grok 4.5: SpaceXAI's coding model challenges Opus and GPT-5.6

    SpaceXAI has released Grok 4.5 – a coding model for $2/6 per million tokens. What the benchmarks show and where caution is advised.

  2. SpaceXAI releases Grok 4.6: cheaper than GPT-5.6 Sol

    The new language model achieves a rating on par with GPT-5.6 Sol according to independent tests, but costs about 60 percent less than the competition.