AI vendor pricing shifts and market share claims
OpenAI introduced a $100/month ChatGPT Pro tier aimed at heavier coding and agent use, while some market reports claim Anthropic has taken substantial share in enterprise code‑generation and agent markets. The industry narrative is fragmenting into segmented pricing and vendor positioning for enterprise use cases. (storyboard18.com, panewslab.com)
OpenAI has added a new $100-a-month ChatGPT Pro tier, splitting its top consumer plan as coding agents become a bigger sales pitch. (openai.com) OpenAI said on April 9 that the new tier offers 5 times the usage of ChatGPT Plus and 10 times the Codex usage of Plus for a limited time, while the existing $200 Pro plan stays in place with 20 times the usage of Plus. OpenAI launched the original ChatGPT Pro plan at $200 a month on December 5, 2024. (openai.com, openai.com, openai.com) Codex is OpenAI’s coding tool inside ChatGPT, and the company said the new $100 plan is meant for “longer, high-effort Codex sessions.” OpenAI’s pricing page now lists Plus at $20 a month and Pro, Business, and Enterprise as separate paid tiers. (openai.com, openai.com) Anthropic is making the same bet from the other side: sell coding help as a seat, not just a model. On August 20, 2025, Anthropic said Team and Enterprise customers could buy premium seats that bundle Claude Code with higher usage and admin controls. (anthropic.com, anthropic.com) Claude Code is Anthropic’s coding agent, which the company says can read a codebase, edit files, run tests, and deliver committed code. In May 2025, Anthropic also said Claude Opus 4 improved coding and let Claude Code run long tasks in the background. (anthropic.com, anthropic.com) The market-share claims driving this story come mostly from private research, not audited financial filings. Menlo Ventures said in its 2025 enterprise survey that Anthropic reached 40% of enterprise large language model spend, up from 24% a year earlier, while OpenAI fell to 21%; in a separate 2025 mid-year update, Menlo said Claude held 42% share in code generation against OpenAI’s 21%. (menlovc.com, menlovc.com) Those figures measure estimated spend and survey responses, not the whole market, and Menlo said its year-end report combined a survey of about 500 United States enterprise decision-makers with its own market model. Anthropic has also published its own usage data showing Claude activity remains concentrated in coding-related work, including software bug fixing. (menlovc.com, anthropic.com) OpenAI and Anthropic are both moving away from one-price-fits-all subscriptions toward usage ladders built around coding and agent work. The next test is whether buyers keep paying for dedicated coding seats, or shift that spending back to broader workplace plans and application software. (openai.com, anthropic.com)
