Latent Space
7/10 signal
[AINews] Codex usage up >10x in 6 months to 7M users, +1M in the past ~day; did Codex overtake Claude Code??
agenticindustry
What happened
Codex has seen explosive growth, increasing its user base by over 10x in six months to reach 7 million users. A recent spike added 1 million users in a single day, coinciding with the launch of GPT-5.6 Sol, suggesting it may have overtaken competitors like Claude Code, which reported around 2 million users in February.
Context
Coding assistants represent one of the most successful applications of agentic AI, with developers rapidly adopting LLMs to augment their workflows. This adoption transforms coding from a manual task to a collaborative process with an AI pair programmer. The competitive landscape is fierce, with major labs like OpenAI (Codex) and Anthropic (Claude Code) vying for dominance. This report's metrics highlight the incredible speed at which these tools are moving from niche products for early adopters to essential infrastructure for the mainstream engineering community, making the developer tool space a key battleground for AI supremacy.
Key points
- Codex's user base has grown by more than 1000% over the last six months, scaling from under 700,000 to 7 million users.
- The launch of a new model, referred to as GPT-5.6 Sol, directly triggered a massive user acquisition event, with Codex adding 1 million new users in approximately one day.
- The growth positions Codex favorably against competitors. For comparison, Claude Code was reported to have roughly 2 million users and a $2.5 billion ARR run rate as of February.
- This rapid scaling demonstrates that coding agents are achieving significant product-market fit and are being integrated into standard developer workflows at an accelerating pace.
- The direct correlation between a new foundation model release and a massive spike in application usage shows how tightly coupled model capability improvements are to user adoption in integrated tools.
What's new
The key novelty is the reported scale and velocity of adoption for a coding agent. While the existence of coding assistants is not new, a 10x growth trajectory to 7 million users in just six months, and a 1-million-user spike in a single day, provides a concrete new benchmark for how quickly AI-native developer tools can become mainstream infrastructure.
Limitations
The analysis is based on a summary, not the primary source, so the exact methodology for calculating user numbers is not detailed. The name 'GPT-5.6 Sol' may be a placeholder or speculative. The comparison to Claude Code's ARR is incomplete without knowing Codex's monetization model or revenue figures.
The take
The developer tool space is the crucible for agentic workflows, and these numbers confirm it's white-hot. A 7 million user count isn't a niche tool; it's a platform shift. This suggests coding agents are becoming as fundamental as the IDE or version control. The most telling detail is the 1M user spike post-model release. It proves that in the AI application layer, user acquisition is a direct function of the underlying model's perceived power. The best model doesn't just win benchmarks; it immediately captures the market.
Signal
Agentic AI is finding its most potent product-market fit within developer tools, scaling faster than in other verticals. The tight feedback loop between foundation model releases and user growth indicates that the application layer's success is, for now, directly tied to the capabilities of the underlying model.
Don't read this site daily. Get it in your inbox.
The daily brief and Sunday deep dive — distilled, scored, and opinionated. For builders only.