SpaceXAI, Meta puts pricing squeeze on Anthropic, OpenAI

OpenAI and Anthropic were already set up to be squeezed by open source large language models (LLM) from China and the US, but now SpaceXAI and Meta are getting in on the fun.

Simply put, the race to the bottom is on and enterprises that were hit by unexpected token costs are likely to cheer. Anthropic and OpenAI aren't going to be thrilled.

Meta released Muse Spark 1.1, its latest model and a big upgrade from the first version of Muse Spark. Meta, which is looking to prove that its Meta Superintelligence Labs unit has momentum, also just launched Muse Image.

According to Meta, Muse Spark 1.1 is comparable to OpenAI's GPT-5.5 and Anthropic's Claude Opus 4.8.

But performance is only half of the story. Enterprises need price for performance. In a post on X, Meta's AI chief Alexandr Wang said Muse Spark 1.1 will be priced at $1.25 per 1 million input tokens and $4.25 per million output tokens.

That pricing, which isn't an introductory rate, undercuts SpaceXAI's Grok 4.5 pricing. SpaceXAI said Grok 4.5 is on par with Claude Opus 4.8. The LLM is priced at $2 per million input tokens and $6 per million output tokens. SpaceXAI claims Grok 4.5 has 2x the token efficiency and can solve task at half the number of steps.

The message from SpaceXAI and Meta is clear: We're cheaper and just as good. We also have the AI infrastructure and funding to do some damage.

Anthropic's Claude Sonnet 5 pricing is $2 per million input tokens through Aug. 31 and then will go up to $3 per million. Output tokens for Claude Sonnet will run $10 per million on an introductory rate then $15 per million.

Claude Opus 4.8 runs $5 per million input tokens and $25 per million output tokens.

OpenAI's GPT 5-5 has a price of $5 per million input tokens and $30 per million output tokens.

Sure, some enterprises will stick with OpenAI and Anthropic, but most companies are already looking at commodity LLMs and open source variants. Given that SpaceXAI and Meta are going to lower the cost curve even more any loyalty to OpenAI and Anthropic is going to be tested.

How we got here:

• DeepSeek's real impact happening now
• Here’s what we learned about AI projects from enterprise buyers so far
• Where’s tokenomics for the rest of us?
• AI inference costs are going to be a big concern: What's the fix?
• Why enterprise AI leaders need to bank on open-source LLMs