OpenAI has introduced GPT-6.1 Sol, a significant upgrade to GPT-6 Sol that offers near-Astra intelligence at one-fifth the cost. This update is particularly relevant for beginners building AI agents, as it provides a more affordable entry point to powerful capabilities without compromising on performance.
Why GPT-6.1 Sol Matters for Beginners
GPT-6.1 Sol is designed to make advanced AI capabilities more accessible. It nearly matches GPT-6 Astra’s performance on tasks like coding, professional work, and computer use, but at a much lower cost. This means you can build and run capable agents without breaking the bank.
Tip
GPT-6.1 Sol’s cached input costs just $0.10 per million tokens—95% less than standard input pricing. This makes it easier to reuse context across requests, a key feature for building efficient agents.
Enhanced Capabilities Across Key Tasks
Coding
GPT-6.1 Sol excels in complex software-engineering tasks. On the DeepSWE v1.1 benchmark, it matches GPT-6 Astra’s performance at one-fifth the cost. This is a substantial improvement over GPT-6 Sol, making it a better choice for coding tasks.
| Model | DeepSWE v1.1 Score | Cost per Task |
|---|---|---|
| GPT-6 Astra | 92.3% | $100 |
| GPT-6.1 Sol | 92.3% | $20 |
| GPT-6 Sol | 85.9% | $25 |
Professional Work
For tasks involving complex PDF documents, GPT-6.1 Sol outperforms Opus 5.5 at less than half the cost. It also approaches GPT-6 Astra’s performance, making it a cost-effective option for professional workflows.
| Model | GDP.pdf Score | Cost per Task |
|---|---|---|
| GPT-6 Astra | 94.7% | $50 |
| GPT-6.1 Sol | 93.2% | $10 |
| Opus 5.5 | 90.5% | $25 |
Computer Use
GPT-6.1 Sol shows significant improvements in computer-use tasks. On OSWorld 2.0’s offline set, it outperforms GPT-6 Sol by seven percentage points at less than half the cost.
| Model | OSWorld 2.0 Score | Cost per Task |
|---|---|---|
| GPT-6 Astra | 89.4% | $70 |
| GPT-6.1 Sol | 87.3% | $10 |
| GPT-6 Sol | 80.3% | $25 |
Scientific Research
For scientific workflows, GPT-6.1 Sol more than doubles GPT-6 Sol’s score at less than half the cost. It’s a cost-effective option for tasks like data analysis and simulation.
| Model | Terminal-Bench Science 0.1 Score | Cost per Task |
|---|---|---|
| GPT-6 Astra | 68.1% | $23.80 |
| GPT-6.1 Sol | 65.4% | $5.47 |
| Opus 5.5 | 63.2% | $23.21 |
Improved Factuality and Safety
GPT-6.1 Sol also improves factual accuracy and safety. It reduces the share of responses containing a factual error by 32% compared to GPT-6 Sol. Additionally, it shows lower failure rates in challenging evaluations, making it a safer choice for agentic tasks.
| Model | Factual Error Rate |
|---|---|
| GPT-6 Astra | 6.5% |
| GPT-6.1 Sol | 7.7% |
| GPT-6 Sol | 11.4% |
Pricing and Availability
GPT-6.1 Sol is available starting today to all Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. Developers can also access it through the OpenAI API as gpt-6.1-sol. Its standard API prices are:
| Token Type | Cost per Million Tokens |
|---|---|
| Input | $2 |
| Cached Input | $0.10 |
| Output | $10 |
In the coming days, OpenAI will also offer GPT-6.1 Sol Ultrafast, with up to 8x faster token generation compared to its standard speed in Codex.
Conclusion
GPT-6.1 Sol represents a significant step forward in making advanced AI capabilities more accessible. Its near-Astra performance at a fraction of the cost makes it an excellent choice for beginners building AI agents. Whether you’re coding, handling professional workflows, or conducting scientific research, GPT-6.1 Sol offers a cost-effective solution without compromising on quality.
Read next
- What OpenAI’s Australia Incident Means for Beginners Building AI
- OpenAI Extends Cyber Defense Tools to Ukraine: What It Means for AI Builders
Want to try all of this hands-on? Start with the free Claude Code from Zero course.
Source
Based on OpenAI’s announcement, “Introducing GPT-6.1 Sol”. Written for people learning to build with these tools.