OpenAI and Anthropic roll out new models on the same day...free access vs. price cuts

[Financial News] OpenAI and Anthropic unveiled new artificial intelligence (AI) models side by side on the same day.
OpenAI has made its next-generation model available even to free users, while Anthropic has introduced a small model with improved performance at sharply reduced prices, fueling the race to bring AI to the masses.
OpenAI said in an official blog post on the 7th (local time) that it was releasing a new GPT-6 model for the at least 1.2 billion people who use ChatGPT every week. The move expands access after the company unveiled its first GPT-6 model to paid users last month.
Paid users can use GPT-6 Sol starting that day, while users on the free and low-cost Go plans can use GPT-6 Luna starting on the 8th.


The key feature in this update is “Intelligent UI.” Depending on the question, ChatGPT can respond with interactive screens—such as charts, diagrams, buttons and forms—instead of text.
When asked, it can also instantly create tools such as a savings calculator, a cost-splitting calculator or a simple game. For questions better suited to plain text, it responds in writing as before.
The feature is available in the general chat interface, but not in business features or the Codex coding tool.
OpenAI described it as “a major change in the way people interact with ChatGPT.”
Responses are also faster. According to OpenAI's internal evaluations, GPT-6 Instant begins responding an average of 44% faster than the previous model to questions that require web searches. It can start answering while it is still thinking.


Rival Anthropic also launched a small AI model, Claude Haiku 5.5, on the same day.
Designed for repetitive, high-volume tasks such as summarization, classification and database lookups, it costs about 75% less on average to run than its predecessor, Haiku 4.5.
The price per 1 million input tokens is $0.10, one-tenth the price of Haiku 4.5 ($1).
Anthropic said Haiku 5.5 is the cheapest, fastest and highest-performing small model it has released to date.
Anthropic said its internal evaluations found that Haiku 5.5 outperformed OpenAI's GPT-6 Luna in key performance comparisons, including computer use and agentic coding.
Anthropic also cut the cache-read price for its mid-sized Sonnet 5.5 model by half, reducing the cost of most agent tasks by about 20%. It also said it would give Max and Team subscribers API credits worth up to $500 a month (about 680,000 won).

[email protected] Lee Seok-woo, International Affairs Correspondent Reporter