Meta CEO Mark Zuckerberg made his highly anticipated return to X on July 26, 2023, to announce the release of Muse Spark 1.1, a new AI coding tool designed to compete in an increasingly crowded market. This marks the first post from Zuckerberg since July 2020, a gap of nearly three years, highlighting the strategic importance of the announcement.
Zuckerberg’s return came with significant implications for the tech industry, particularly as it pertains to the ongoing AI race. His X post revealed that Meta is positioning Muse Spark 1.1 as a “strong agentic and coding model at a very low price,” aiming to attract developers and enterprises looking for cost-effective alternatives to existing AI coding tools.
The AI coding space is already dominated by models like Claude Code from Anthropic and Codex from OpenAI. To stand out, Meta has set competitive pricing for Muse Spark 1.1. Every new API account will start with $20 (approximately Rs 1,900) in free credits, followed by charges of $1.25 per million input tokens and $4.25 per million output tokens. In comparison, the recently released Grok 4.5 from Elon Musk’s SpaceXAI is priced at $2 (roughly Rs 190) per million input tokens and $6 (approximately Rs 570) per million output tokens.
Meta claims that Muse Spark 1.1 outperforms other models like Anthropic’s Opus 4.8 and OpenAI’s GPT-5.5 in certain tasks, as evidenced by benchmark comparisons shared by Zuckerberg on X. However, these benchmarks did not include Anthropic’s Mythos and Fable models or OpenAI’s GPT-5.6.
Developers can access Muse Spark 1.1 through the Meta Model API. Pricing is structured to offer a cost advantage over competitors like Anthropic, which charges $5 (approximately Rs 480) per million input tokens and $25 (roughly Rs 2390) per million output tokens, or OpenAI’s GPT-5.6 Terra, priced at $2.5 (approximately Rs 240) per million input tokens and $15 (roughly Rs 1430) per million output tokens.
Zuckerberg’s announcement came amidst a flurry of activity in the AI space. On Tuesday, Meta released Muse Image, its new image-generation model, through the Meta AI app, WhatsApp, and Instagram. Additionally, OpenAI launched GPT-5.6 and its work AI agent, ChatGPT Work, marking a shift in Meta’s strategy from open-source releases to proprietary models.
In response to Zuckerberg’s announcement, Elon Musk chimed in on X with a playful “Jinx,” underscoring the competitive nature of this tech showdown.
Alexandr Wang, head of Meta’s Superintelligence Labs (MSL), emphasized that while Meta is now charging for access to its AI tools, it remains committed to open-source initiatives. An open-source variant of Muse Spark is in development but has not yet been released. Wang also mentioned the training of a more powerful model codenamed Watermelon.
With this pricing strategy and competitive capabilities, Meta aims to reclaim ground in the rapidly evolving AI landscape.
Thinking about building an AI product?
Get in Touch