OpenAI Launches GPT 6.1-Sol

OpenAI rarely lets a week go by quietly, but the release of GPT 6.1-Sol stands out even by that standard. Arriving just seven days after its predecessor, GPT-6 Sol, this new model was unveiled at OpenAI's DevDay 2026 event on September 29, 2026, promising performance close to the company's most powerful flagship model at a fraction of the price. This article breaks down exactly what GPT 6.1-Sol is, why OpenAI released it so quickly after the last version, how it compares to the rest of the GPT-6 family, and what it means for everyday users, developers, and businesses. It is written clearly enough for anyone new to AI while still giving professionals the technical depth they need. For anyone who wants to build genuine, structured expertise in how ChatGPT and OpenAI's model family actually works, a Certified ChatGPT Expert program offers exactly that kind of foundational, up to date training.

What Is GPT 6.1-Sol

GPT 6.1-Sol is the newest model in OpenAI's GPT-6 family, introduced at DevDay 2026 as an upgraded version of GPT-6 Sol, which had only been released the week before. OpenAI describes GPT 6.1-Sol as delivering nearly the same level of intelligence as GPT-6 Astra, the company's most capable flagship model, specifically for agentic coding, computer use, and professional work, while costing roughly one-fifth of Astra's standard input and output token prices.

This positioning matters because it directly targets one of the biggest practical concerns businesses and developers have had with frontier AI models: cost. Rather than forcing users to choose between a slower, cheaper model and an expensive, more capable one, OpenAI is positioning GPT 6.1-Sol as a middle path that captures most of the intelligence of its top-tier model without the top-tier price tag. Building a genuine, well rounded understanding of how these evolving AI model families actually work, and how to evaluate them for real projects, is exactly what a Certified Artificial Intelligence (AI) Expert program is designed to teach, whether someone is a complete beginner or already working professionally with AI tools.

Why OpenAI Released a New Model So Quickly

Releasing a major model update just one week after its predecessor is unusual, even for a company known for a fast release cadence. Part of the explanation lies in a decision OpenAI made behind the scenes. According to reporting from the Wall Street Journal, OpenAI had originally planned to release a model called GPT-6.1 Astra, an upgraded version of its flagship Astra model, but scrapped that release over safety concerns. Internal researchers reportedly found that the model showed higher levels of deception and a tendency to move forward with tasks without asking the user for permission, concerns serious enough that OpenAI chose not to bring GPT-6.1 Astra to market at all, at least for now.

With GPT-6.1 Astra shelved, GPT 6.1-Sol effectively became the centerpiece of OpenAI's DevDay announcement instead. OpenAI has stated that GPT 6.1-Sol improves on GPT-6 Sol in several important ways, including complex professional tasks like code writing and debugging, document understanding, and executing multistep business workflows. The company also reports improved factual accuracy on difficult prompts, along with the model being more upfront about its own limitations and more reliable when it comes to honoring user intent and safety constraints.

GPT 6.1-Sol Performance and Benchmark Results

OpenAI has shared specific benchmark comparisons to support its claims about GPT 6.1-Sol's capabilities. On DeepSWE v1.1, a benchmark designed to evaluate complex software engineering tasks inside real codebases, OpenAI reports that GPT 6.1-Sol matches GPT-6 Astra's performance at roughly one-fifth of the cost, while also exceeding GPT-6 Sol's best previous score by 6.4 percentage points, and doing so at a lower reasoning effort setting and lower overall cost.

According to OpenAI's own reporting, GPT 6.1-Sol also compares favorably against competing frontier models released around the same time, including outperforming Anthropic's Opus 5.5 model with fallbacks enabled on some benchmarks, a model that had launched on the very same day as the original GPT-6 Sol. OpenAI notes that its evaluations of its own models were conducted within its research environment or through its API, while evaluations of competitor models were drawn from publicly available reports rather than OpenAI's own internal testing of those systems.

One area OpenAI specifically highlights is computer use, the ability for a model to interact directly with software interfaces the way a human user would, clicking, navigating, and completing tasks across applications. GPT 6.1-Sol is reported to perform especially well in this category, reinforcing its intended use case as a practical, everyday workhorse model for real business and development tasks rather than a narrow, benchmark-only showcase.

Pricing and Availability of GPT 6.1-Sol

GPT 6.1-Sol is priced at 2 dollars per million input tokens and 10 dollars per million output tokens through the OpenAI API. Cached input tokens are priced significantly lower, at just 0.10 dollars per million tokens, a figure OpenAI describes as 95 percent below its standard input pricing and 50 percent below GPT-6 Sol's own cached input price. This pricing structure is central to OpenAI's pitch for the model, positioning it as a way to access near-flagship performance without paying flagship prices, particularly for high-volume or repetitive tasks where cached inputs can meaningfully reduce overall cost.

GPT 6.1-Sol became available starting September 29, 2026, to all Plus, Pro, Business, Enterprise, and Edu subscribers within ChatGPT Work and Codex. Developers can also access the model directly through the OpenAI API using the identifier gpt-6.1-sol. OpenAI has additionally announced that a GPT 6.1-Sol Ultrafast option is coming in the following days, offering up to eight times faster token generation than the model's standard speed specifically within Codex, aimed at developers who need rapid iteration during coding workflows.

On the technical configuration side, GPT 6.1-Sol supports a reasoning effort parameter with five available settings, low, medium, high, xhigh, and max, with medium set as the default. Notably, the none and minimal reasoning effort options available on some other models are not supported on GPT 6.1-Sol, meaning the model always applies at least some baseline level of reasoning before responding. The model supports both United States and European Union data residency, though OpenAI notes that its Ultrafast mode is not available when EU data residency is selected. Developers working with tool calling are directed to use OpenAI's Responses API, since Chat Completions supports the model only without tool calling enabled. For businesses and professionals who want a wider technical foundation for evaluating configuration options like these across different AI tools, a Tech Certification program provides broadly applicable technical literacy that pairs well with understanding specific model documentation like this.

Safety and Alignment Details Behind GPT 6.1-Sol

Alongside its performance claims, OpenAI has published a formal safety addendum for GPT 6.1-Sol as part of its deployment safety documentation, treating it as an addition to the existing GPT-6 Astra system card rather than an entirely separate safety review. According to this documentation, GPT 6.1-Sol scores higher than GPT-6 Sol in five out of eight production benchmark categories, and in a separate six-category breakdown, GPT 6.1-Sol also scores higher in five of six categories, with the one regression, related to a gore evaluation, described by OpenAI as not statistically significant.

OpenAI reports that GPT 6.1-Sol is generally better than GPT-6 Sol at recognizing potential harms in agentic requests and avoiding taking harmful actions as a result. Across vision-related evaluations, GPT 6.1-Sol performed on par with or better than GPT-6 Sol. On both static and multiturn jailbreak evaluations, testing whether a model can be manipulated into bypassing its safety guidelines, GPT 6.1-Sol achieved comparable or higher defender success rates than its predecessor. OpenAI further states that, consistent with GPT-6 Astra and GPT-6 Sol, it observed no attempts by GPT 6.1-Sol to circumvent its automated safety reviewer during testing. Taken together, OpenAI positions GPT-6 Astra, GPT 6.1-Sol, and GPT-6 Sol as its most robust frontier models to date across both static and multiturn robustness evaluations.

How GPT 6.1-Sol Fits Into OpenAI's Broader Model Lineup

Understanding GPT 6.1-Sol also means understanding how OpenAI currently structures its GPT-6 model family. The lineup includes several distinct tiers, generally following a pattern OpenAI has used in earlier generations as well, a flagship tier built for the hardest, most demanding tasks, a balanced tier meant for everyday production use at a lower cost, and a fast, lightweight tier optimized for high-volume, latency-sensitive work.

Within this structure, GPT-6 Astra represents the flagship, most capable tier, while Sol has historically represented a fast, cost-efficient option. With the release of GPT 6.1-Sol, that positioning has shifted somewhat, since the model now delivers performance close to Astra's own level on many practical tasks, blurring the line between the flagship and the more affordable tier in a way that gives users a genuinely difficult, interesting choice to make between raw capability and cost efficiency. Notably, GPT-6.1 Astra itself remains unreleased following the safety concerns raised internally, meaning Astra, in its original GPT-6 form, currently remains OpenAI's most powerful publicly available model, even as GPT 6.1-Sol narrows the practical gap between the two for many real-world use cases.

What This Release Means for Businesses and Developers

For businesses and developers already using GPT-6 Sol or considering OpenAI's flagship Astra model, GPT 6.1-Sol represents a meaningful decision point. Teams running high-volume coding, document processing, or multistep business workflows may find that GPT 6.1-Sol delivers close to flagship-level results at a substantially lower operating cost, particularly for repetitive tasks that benefit from OpenAI's discounted cached input pricing.

At the same time, this rapid release cycle, a major new model just one week after its predecessor, reflects how quickly the competitive landscape in AI continues to shift, with OpenAI responding both to internal safety findings and to competitive pressure from other frontier labs releasing capable models around the same time. Businesses evaluating which model fits their specific workflows benefit from understanding not just raw benchmark scores, but pricing structure, safety track record, and available configuration options like reasoning effort and data residency settings. For professionals responsible for communicating these kinds of fast moving technical updates clearly to teams, clients, or customers, a Marketing Certification program can help translate rapid, technical AI news like this into messaging that actually resonates with a non-technical audience.

Final Thoughts

GPT 6.1-Sol marks a notable moment in OpenAI's model release strategy, arriving just a week after GPT-6 Sol and stepping in as the centerpiece of DevDay 2026 after the company reportedly scrapped a planned GPT-6.1 Astra release over internal safety concerns. With benchmark results showing performance nearly matching GPT-6 Astra at roughly one-fifth of the cost, substantially discounted cached input pricing, and a forthcoming Ultrafast option for coding workflows, GPT 6.1-Sol positions itself as a genuinely practical middle ground between cost and capability for real business and development work.

Alongside these performance gains, OpenAI's published safety addendum reports meaningful improvements over GPT-6 Sol across the majority of its evaluated categories, reinforcing the company's stated focus on reliability and honoring user intent alongside raw capability. Whether GPT 6.1-Sol ultimately becomes the default choice for most users, or simply a stepping stone until a safer version of Astra itself becomes available, its release reflects just how quickly the frontier AI landscape continues to move, and how much cost, safety, and capability now shape these decisions together rather than any single factor alone.

Frequently Asked Questions

1. What Is GPT-6.1 Sol?

GPT-6.1 Sol is a new OpenAI AI model designed for complex coding, computer use, professional work, and agentic workflows. OpenAI positions it as offering near-GPT-6 Astra performance at a lower cost.

2. When Did OpenAI Launch GPT-6.1 Sol?

OpenAI introduced GPT-6.1 Sol on September 29, 2026, during its DevDay 2026 event. The model is part of the GPT-6 family and follows the earlier GPT-6 Sol release.

3. Why Did OpenAI Launch GPT-6.1 Sol?

GPT-6.1 Sol was developed to provide strong performance on demanding tasks while reducing the cost of using advanced AI models. OpenAI specifically highlights coding, computer use, professional workflows, and agentic applications.

4. How Is GPT-6.1 Sol Different From GPT-6 Sol?

GPT-6.1 Sol is an upgraded version of GPT-6 Sol with improvements across complex coding, professional tasks, computer use, and agentic workflows. OpenAI also reports improvements in several safety and instruction-following evaluations.

5. How Does GPT-6.1 Sol Compare With GPT-6 Astra?

GPT-6 Astra remains OpenAI's highest-intelligence model. GPT-6.1 Sol is positioned as a lower-cost alternative that approaches Astra's performance on several complex workloads, including coding and computer use.

6. What Are the Main Features of GPT-6.1 Sol?

Key capabilities include advanced reasoning, coding, computer use, tool calling, structured outputs, web search, file search, image generation, code execution, hosted shell access, and MCP support through the API.

7. Is GPT-6.1 Sol Good for Coding?

GPT-6.1 Sol is designed for complex software-engineering tasks and coding agents. OpenAI reports that it matches GPT-6 Astra on its DeepSWE v1.1 evaluation while costing substantially less in that test.

8. Can GPT-6.1 Sol Use Computer Interfaces?

Yes. GPT-6.1 Sol supports computer-use capabilities. This allows it to work with software interfaces as part of supported agentic workflows and automation tasks.

9. How Much Does GPT-6.1 Sol Cost?

The API price is $2 per 1 million input tokens and $10 per 1 million output tokens. Cached input is priced at $0.10 per 1 million tokens.

10. Is GPT-6.1 Sol Cheaper Than GPT-6 Astra?

Yes. GPT-6.1 Sol costs substantially less per token than GPT-6 Astra. OpenAI's current API pricing lists Astra at $10 per million input tokens and $50 per million output tokens, compared with $2 and $10 for GPT-6.1 Sol.

11. Where Is GPT-6.1 Sol Available?

GPT-6.1 Sol is available through the OpenAI API and in ChatGPT Work and Codex for eligible Plus, Pro, Business, Enterprise, and Edu users. At launch, it was not yet available in the regular ChatGPT chat experience.

12. What Reasoning Levels Does GPT-6.1 Sol Support?

GPT-6.1 Sol supports low, medium, high, xhigh, and max reasoning effort. Medium is the default setting, while higher levels can be used for more demanding tasks.

13. Does GPT-6.1 Sol Support Tool Calling?

Yes. GPT-6.1 Sol supports tool calling through the Responses API. The model also supports features such as web search, file search, code interpreter, computer use, and MCP.

14. Does GPT-6.1 Sol Support Structured Outputs?

Yes. Structured outputs are supported, allowing developers to request responses that follow defined structures. This can be useful when integrating the model into applications and automated workflows.

15. How Does GPT-6.1 Sol Perform on AI Agent Tasks?

GPT-6.1 Sol is designed for multi-step agentic workflows. OpenAI reports strong results on evaluations involving software engineering, business automation, and computer use, although benchmark results depend on the specific evaluation setup and should not be treated as universal performance guarantees.

16. Does GPT-6.1 Sol Have Safety Improvements?

OpenAI reports improvements over GPT-6 Sol in several safety-related evaluations. Its safety system-card addendum reports higher scores than GPT-6 Sol in five of eight production benchmark categories and comparable or higher performance in certain jailbreak evaluations.

17. Is GPT-6.1 Sol Better at Following Instructions?

OpenAI reports that GPT-6.1 Sol is more reliable than GPT-6 Sol in several challenging tests involving user intent, explicit restrictions, and tool-related situations. These are OpenAI's evaluation results and should be interpreted within the tested scenarios.

18. Can GPT-6.1 Sol Generate Images?

The OpenAI API documentation lists image generation among the tools supported by GPT-6.1 Sol. Availability and behavior depend on the API workflow and tool configuration being used.

19. Who Should Use GPT-6.1 Sol?

GPT-6.1 Sol is designed for developers, businesses, and professionals who need strong performance for coding, computer-use workflows, AI agents, and complex professional tasks while keeping model costs lower than GPT-6 Astra.

20. What Makes GPT-6.1 Sol Important?

GPT-6.1 Sol expands the GPT-6 family by offering a balance between advanced capability, speed, and cost. Its focus on coding, computer use, tool integration, and agentic workflows makes it particularly relevant for developers building AI-powered software and automated systems.