OpenAI Unveils GPT-6.2 Astra and Sol Ahead of Anthropic IPO

Original title: Anthropic 上市前夕 OpenAI 正式发布 GPT-6.2 Astra 和 Sol:部分领域超越 Fable 5.2 且成本大降,金融、科研刷新纪录
88AI Score

OpenAI released GPT-6.2 Astra and GPT-6.2 Sol one day before Anthropic's IPO pricing, targeting professional work, coding, computer use and multi-step agentic tasks while cutting output token costs. OpenAI reports Astra scoring 5.5 points above Claude Fable 5.2 on Agents' Last Exam at just 5% of the cost, plus a new high on Terminal-Bench-Science 0.1. The models ship first to ChatGPT Work, Codex and the API, not the standard chat mode.

Full Report

IT之家 November 18 news — just one day before Anthropic's IPO pricing, OpenAI suddenly delivered a "surprise" with no warning.

OpenAI today officially released its latest GPT-6.2 Astra and GPT-6.2 Sol models, with a focus on improving performance in professional work, coding, computer operation, and multi-step agentic tasks, while reducing the number of output tokens required to complete tasks. In most tests, they are on par with Anthropic's flagship model.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

OpenAI introduced that GPT-6.2 Astra is a flagship model for complex professional projects, further improved in areas such as professional work, scientific research, software development, and visual design. The model can more accurately follow user-provided presentation templates, improve the output quality of slides, spreadsheets, and documents, while also enhancing front-end development, website design, game development, and 3D design capabilities.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

In Agents' Last Exam, which tests multi-step professional work, GPT-6.2 Astra scores 5.2 percentage points higher than the previous generation GPT-6 Astra, while the cost per individual task is less than one-third of the latter's; at the same time, it scores 5.5 points higher than Claude Fable 5.2, while costing only 5% of it.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

In addition, GPT-6.2 Astra achieved a new high in the Terminal-Bench-Science 0.1 test (this test is designed to evaluate a model's ability to perform scientific work in a Terminal environment). The new model's score is also 9 percentage points higher than GPT-6 Astra, with task costs likewise less than one-third.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

OpenAI also stated that GPT-6.2 Astra set a new high in the Vals Science Index evaluation as well (this test requires the model to interpret real biological data and run experiments to infer how a simulated system works).

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

In Artificial Analysis's AA-3D evaluation, the model ranked first with 1318 Elo, higher than Claude Fable 5.2's 1271 Elo and GPT-6.2 Sol's 1259 Elo.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

The other model, GPT-6.2 Sol, is positioned for everyday work and is a model that balances efficiency and general-purpose capability. Compared with GPT-6.1 Sol, the new model has improved in agentic workflows, code writing, and multi-step task handling, can complete tasks with fewer tokens, and provides more natural text expression.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

OpenAI said that in Cognition's FrontierCode 2.0 evaluation, GPT-6.2 Sol's score is less than 0.5 points apart from GPT-6.2 Astra, while task cost is about one-fifth of the latter's. In the LifeVal everyday task evaluation, GPT-6.2 Sol's score exceeds GPT-6 Astra, with task cost about one-seventh of the latter's.

On the eve of Anthropic's IPO, OpenAI drops dual models GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some areas with sharply lower costs, setting new records in finance and scientific research

In terms of pricing, the standard price for the GPT-6.2 Astra API is $8 per million input tokens, $0.4 per million cached input tokens, and $40 per million output tokens. Compared with GPT-6 Astra, the new model's input and output prices are both reduced by 20%, and the cached input price is reduced by 60%.

On the eve of Anthropic's IPO, OpenAI drops two models out of the blue, GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some domains with sharply lower costs, setting new records in finance and scientific research

GPT-6.2 Sol keeps the API pricing of GPT-6.1 Sol: $2 per million input tokens, $0.1 per million cached input tokens, and $10 per million output tokens.

OpenAI says that across multiple benchmarks, both models achieve performance close to or even beyond other frontier models at a cost far lower than theirs.

On the eve of Anthropic's IPO, OpenAI drops two models out of the blue, GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some domains with sharply lower costs, setting new records in finance and scientific research

In terms of availability, OpenAI says both models will be available starting November 17 to Plus, Pro, Enterprise, and Edu users of ChatGPT Work and Codex.

Pro 500 users can use the Ultrafast mode of both models, while Pro 200 users gain access to the Ultrafast mode of GPT-6.2 Sol for the first time. Eligible Enterprise and Edu users also need a workspace administrator to enable the relevant features.

On the eve of Anthropic's IPO, OpenAI drops two models out of the blue, GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some domains with sharply lower costs, setting new records in finance and scientific research

Developers can also call gpt-6.2-astra and gpt-6.2-sol through the OpenAI API, and both models offer Ultrafast mode at launch.

On the eve of Anthropic's IPO, OpenAI drops two models out of the blue, GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some domains with sharply lower costs, setting new records in finance and scientific research

According to OpenAI's official release notes, GPT-6.2 Astra and GPT-6.2 Sol are mainly aimed at ChatGPT Work, Codex, and the API this time, and have not yet been added to the regular ChatGPT chat mode.

On the eve of Anthropic's IPO, OpenAI drops two models out of the blue, GPT-6.2 Astra and Sol: surpassing Fable 5.2 in some domains with sharply lower costs, setting new records in finance and scientific research