OpenAI and Anthropic push models
- OpenAI released GPT-5.5 on April 23, and Anthropic launched Claude Opus 4.7 on April 16, compressing two flagship model upgrades into eight days. - GPT-5.5 reached ChatGPT on April 23 and the API on April 24; Anthropic kept Opus 4.7 pricing at $5 in, $25 out. - The pace follows OpenAI’s March GPT-5.4 launch and frequent model-note updates, forcing developers to plan for faster swaps. (openai.com)
OpenAI shipped GPT-5.5 on April 23, 2026, one week after Anthropic made Claude Opus 4.7 generally available on April 16. (openai.com) (anthropic.com) Both companies framed the releases around harder office and coding work, not chatbot small talk. OpenAI said GPT-5.5 is built for code, web research, spreadsheets, documents, and software tools; Anthropic said Opus 4.7 improves advanced software engineering and long-running tasks. (openai.com) (anthropic.com) The timing is the story. Anthropic’s model landed first on April 16, then OpenAI followed on April 23, with OpenAI adding API access for GPT-5.5 and GPT-5.5 Pro on April 24. (anthropic.com) (openai.com) These are the systems developers wire into products, so each release can force changes in prompts, routing rules, latency budgets, safety settings, and fallback logic. OpenAI’s own help pages show that pattern: GPT-5.4 arrived on March 5, GPT-5.4 mini rolled into ChatGPT on March 18, and GPT-5.5 followed on April 23. (openai.com 1) (openai.com 2) A frontier model is the most capable system a lab is willing to ship at that moment. Companies expose those models through application programming interfaces, or APIs, so other apps can call them the way a website calls a payment processor or a map. (openai.com) (anthropic.com) That means product teams are not just choosing “the best model.” They are choosing how tightly to couple their software to one vendor’s model names, tool behavior, context limits, and pricing. Anthropic said Opus 4.7 keeps the same price as Opus 4.6 at $5 per million input tokens and $25 per million output tokens. (anthropic.com) OpenAI’s release notes point in the same direction on the product side. On April 28, ChatGPT moved model selection into the composer and shifted thinking controls into the model picker, another sign that model choice is becoming a more active part of using the product. (openai.com) The benchmark claims also show why teams keep retesting. OpenAI’s GPT-5.5 launch page compares the model against GPT-5.4, Claude Opus 4.7, and Gemini 3.1 Pro on coding, browsing, math, and tool-use evaluations, while Anthropic says Opus 4.7 improves on Opus 4.6 and adds stronger vision. (openai.com) (anthropic.com) Anthropic paired Opus 4.7 with new cyber safeguards and a Cyber Verification Program for legitimate security work. OpenAI said it tested GPT-5.5 with nearly 200 early-access partners and updated its system card on April 24 with added API safeguards. (anthropic.com) (openai.com 1) (openai.com 2) The result is a market where flagship models now change on a weekly cadence, not a yearly one. In April alone, Anthropic launched Opus 4.7 on April 16 and OpenAI launched GPT-5.5 on April 23, then expanded access a day later. (anthropic.com) (openai.com)
