What a coincidence: three AI power players released new models to the public on the same day.
What happened: OpenAI released GPT 5.6, its latest highly advanced model (reportedly after getting White House approval); Meta dropped Muse Spark 1.1, marking the first time the company has offered a paid tier; and SpaceXAI (formerly xAI) unveiled Grok 4.5. Each company has said something along the lines of ‘it’s our strongest one yet’ about their model.
While GPT 5.6 is a general-purpose model, the others are focused on coding and other agentic tasks. Meta and SpaceXAI are both looking to compete with OpenAI and Anthropic on that front.
Zoom in: Meta and SpaceXAI are betting they can win over customers with lower costs and comparable performance. Meta said Muse Spark 1.1 will have “very aggressive and attractive” pricing — part of a long-term wager by CEO Mark Zuckerberg that all enterprise models will eventually do basically the same thing and perform with negligible differences.
Meanwhile, Grok 4.5 is priced at US$2 per million input tokens and $6 per million output tokens. For comparison, Anthropic's Claude Opus 4.8 — a very popular enterprise AI model — is priced at $5 per million input tokens and $25 per million output tokens.
Why it matters: More companies are getting fed up with how much it costs to use enterprise models, going from tokenmaxxing to tokenminning, and could soon feel pressured to cut AI budgets — especially if they aren’t seeing a marked improvement in productivity. With a potential price war looming, being the cheapest option might be a solid strategy.—QH
