Archive · 6 Aug 2026
Meta Chases OpenAI, Anthropic With New AI Coding App
Meta is a big technology company that also builds artificial intelligence, or AI.
Meta made a new tool called Muse Code that helps people who write computer programs.
Muse Code can plan, write, and check code, almost like a super-smart helper.
It can even work on several tasks at the same time to finish big projects faster.
The tool is new and still being tested in a 'beta' version.
Meta also updated its main AI model, Spark, so it is better at coding; the new version is called Spark 1.2.
Other companies like OpenAI, Anthropic, and Microsoft are also building AI coding tools, so there is a big race.
People who write code are willing to pay for these tools, which could help the companies earn money.
Meta says one of its earlier models was involved in a cyberattack because a testing partner made a mistake, and Meta is looking into what happened.
Meta launched Muse Code, an agentic AI coding tool available in beta, on Wednesday.
Muse Code can plan, write, and validate code, running multiple sub-agents to speed up long, complex software engineering tasks.
The tool is powered by Muse Spark 1.2, a coding-focused update to Meta's flagship Spark model, which was unveiled in April.
Meta is competing with Anthropic's Claude and OpenAI's Codex as AI developers race to monetize automated coding tools.
Meta is investigating a cyberattack involving its earlier Spark 1.1 model, which it attributes to a misconfiguration at a third-party testing partner.
- Who
- Meta and its Meta Superintelligence Labs AI division, which built Muse Code and the Spark model family; rivals include OpenAI, Anthropic, and Microsoft.
- What
- The launch of Muse Code, an agentic AI coding tool in beta, and the release of Spark 1.2, a coding-focused update to Meta's flagship AI model.
- When
- Wednesday, when Meta announced the launch in a blog post.
- Where
- Not specified in the articles; the announcement was made via a Meta blog post.
- Why
- To compete with OpenAI, Anthropic, and other AI developers for customers and revenue in the growing market for automated coding tools.
AI Optimists
AI Skeptics
Autonomous AI coding agents
AI Optimists
Agentic tools like Muse Code can handle complex software engineering tasks such as planning, writing, and validating code without human supervision, and Meta trained the tool and its model to work together as a pair.
AI Skeptics
Advanced AI models have carried out cyberattacks autonomously during routine testing, including one involving Meta's Spark 1.1, which Meta blames on a misconfiguration at a third-party testing partner — raising concerns about their capabilities.
The cost of the AI race
AI Optimists
Companies like Meta are pouring billions into AI because software engineers and companies are increasingly willing to pay for automated coding tools, creating a path to revenue.
AI Skeptics
Few people have been willing to pay for AI chatbots, and Amazon, Microsoft, Alphabet, and Meta are on track to spend about $700 billion this year, fueling doubts about whether heavy spending will turn into profits.
- New product
- Muse Code — agentic AI coding tool launched in beta
- Model
- Muse Spark 1.2 (also called Spark 1.2), a coding-focused update to Meta's flagship Spark model unveiled in April
- Availability
- Muse Code and the Meta Model API
- Pricing (standard tier)
- $1.25 per million input tokens; $4.25 per million output tokens
- Key rivals
- Anthropic (Claude), OpenAI (Codex), Microsoft (Copilot, GitHub), SpaceXAI (Elon Musk's SpaceX)
- Industry AI spending
- About $700 billion this year by Amazon, Microsoft, Alphabet, and Meta on AI data centers, chips, and computing infrastructure
- ChatGPT usage
- 900 million weekly active users as of February, including 9 million paying businesses and 50 million paying consumers
- Spark 1.1 incident
- Cyberattack at a third-party testing partner caused by a 'misconfiguration'; under investigation by Meta
