The latest
The contest between OpenAI and Anthropic is no longer defined only by user counts or the spectacle of new model launches. It is increasingly about who owns the workplace tool that developers and companies are willing to pay for. OpenAI bet that ChatGPT could become a broad consumer gateway to AI. Anthropic, meanwhile, treated coding as a clearer commercial opportunity. Its Claude Code product gained traction quickly, helping the company overtake its rival in revenue growth and valuation, according to The Wall Street Journal.
Details
- A different bet: OpenAI made ChatGPT’s growth central to its strategy while continuing to spend on a broad mix of projects, including video generation, devices and chips.
- The coding gap: The company had reasoning models that were theoretically well suited to writing code, but they were first trained on narrowly defined programming problems rather than the messier work software teams face inside companies.
- Anthropic’s approach: Anthropic designed Sonnet 3.7 around real-world tasks, then built on that focus with Claude Code, which drew software engineers looking for a practical coding assistant.
- Codex struggles: OpenAI gave employees early access to Codex, but adoption fell short of expectations. After its public launch in May, some developers found it slower and more cumbersome than its rival.
- Internal pressure: The problem unfolded as OpenAI dealt with Meta’s recruiting push, tensions with Microsoft and computing costs tied to projects that did not deliver as expected, including Sora.
- The financial signal: Anthropic’s lead in revenue growth and valuation has raised concern among some OpenAI investors given the company’s high cash burn. Executives are also weighing enterprise pricing for Codex, a move that could win market share while putting pressure on margins.
Between the lines
The issue is not simply that OpenAI was late with a coding product. It misread where the market was shifting. Consumer AI created attention, but specialised productivity tools are beginning to generate the recurring revenue and durable corporate relationships that matter most. Anthropic’s success gives it a stronger narrative ahead of a potential IPO and forces OpenAI to show that its expanding product range is an advantage rather than a distraction.
What to watch
OpenAI is betting on GPT-5.6 Sol and a tighter integration of Codex, ChatGPT and a web browser to win back developers and knowledge workers. The real test will be whether that momentum turns into recurring enterprise contracts without discounts and free computing eroding the economics of the business ahead of a public listing.
