SpaceXAI、競合他社より低価格な「Grok 4.6」を公開し長期間エージェントに注力

2 Min Read

Joe Raedle via Getty Images

As the AI pricing war ramps up, Elon Musk’s SpaceXAI is aiming to prove to enterprises that it offers the most cost-efficient model with the release of Grok 4.6.

Introduced this week, a little over a month after Grok 4.5, Grok 4.6 focuses more on long-running agents. It was trained on various agentic reinforcement learning tasks, including knowledge work, general coding and domain-specific environments, SpaceXAI said. The model is effective at turning a broad product idea into a working first version and at assessing its own work, according to the vendor.

While Grok 4.6 competes with frontier models such as Anthropic Claude Code and OpenAI’s GPT-5.6 in supporting long-running agents, it is priced lower than the other models. With the pricing, SpaceXAI appears to be appealing to enterprises feeling the crushing weight of using pricier AI models.

“It’s really more about AI economics than anything else,” said Sid Nag, founder and chief research officer of independent analyst firm Tekonyx. “There's a pricing war going on in this industry.”

Related:Euro Vibe Coding Standout Lovable Now Valued at $13.3B

AI Pricing War

The price-jockeying in the AI market is evident, with model providers offering cheaper models in recent months. For example, OpenAI’s latest model is cheaper than previous releases, and the same can be said for Anthropic's recent releases. Nevertheless, Grok 4.6 is still less expensive than models from its two more established rivals, which have much larger customer bases.

Grok 4.6 costs $2 per million input tokens and $6 per million output tokens. By comparison, GPT-5.6 Sol costs $5 per million input tokens and $15 per million output tokens. Meanwhile, Claude Opus 5 goes for $5 per million input tokens and $25 per million output tokens.

“Grok is addressing the tokenomics aspect of the market,” Nag said, referring to the fast-developing practice of economizing on token usage.

Because Grok 4.6 is connected to the AI coding platform Cursor, which SpaceXAI acquired in June, this can also be attractive to enterprise developers. Moreover, the focus on long-running agents could lead to more productivity for enterprises. Nag said that when enterprises need to increase their input and output token counts to boost productivity, the model that lets them do so at lower token pricing could be the one they choose.

“So that boils down to pricing,” he said.

Some Challenges

However, SpaceXAI still faces some challenges. For example, cheaper tokens do not mean cheap AI, Nag said. Enterprises also need to consider other costs, such as APIs and other costs involved in the AI and agentic AI process, such as inference and caching, which are not included in token pricing.

Related:Mistral Aims to Build 1GB of Compute Capacity by 2030

Nag added that the vendor will also have to ensure its governance policy aligns with enterprise environments.

Moreover, SpaceXAI might need to consider its model lifecycle and release schedule so that enterprises can plan how to implement them. The vendor just released Grok 4.6, and Musk said this week the vendor plans to release Grok 4.7 soon.

About the Author

News Writer, AI Business

Esther Shittu has covered AI technologies and industry trends since 2021. As co-host of the Targeting AI podcast, she talks with experts, thought leaders and practitioners exploring critical AI developments. Before AI Business, she wrote for SearchEnterpriseAI, the New York Daily News, Bklyner and the Brooklyn Daily Eagle. When she's not diving deep into the world of AI, she spends her time on passion projects and raising her three daughters.