Sold by: OpenAI
Deployed on AWS
AWS Free Tier
OpenAI GPT-6.1 Sol brings intelligence approaching GPT-6 Astra to everyday work on Amazon Bedrock. Build and scale capable agents for coding, computer use, and professional workflows at Sol pricing.
3.5
Overview
OpenAI GPT-6.1 Sol is available on Amazon Bedrock, bringing stronger reasoning to agentic coding, computer use, and professional work. A major upgrade to GPT-6 Sol, it approaches GPT-6 Astra on several demanding evaluations, with standard input and output token prices roughly one fifth those of GPT-6 Astra. Investigate complex codebases, build features, debug issues, and iterate on solutions with more room to scale. Use GPT-6.1 Sol with Codex on Amazon Bedrock across investigation, implementation, and testing. Turn complex information into action with agents that analyze documents, use tools, and complete workflows across business systems. Build internal tools that synthesize documents and applications that evaluate multiple inputs. Explicit prompt caching on Amazon Bedrock supports workloads that reuse context across repeated requests. Run these workloads with the performance, scalability, security, and reliability of Amazon Bedrock, using established AWS controls to govern access and audit model activity.
Highlights
- Intelligence approaching GPT-6 Astra on several demanding evaluations at Sol pricing, with standard input and output token prices roughly one fifth those of GPT-6 Astra.
- Stronger agentic coding, computer use, and professional work. Investigate codebases, build features, debug issues, analyze complex documents, and complete workflows across business tools.
- Build and scale agents on Amazon Bedrock with explicit prompt caching, production infrastructure, and established AWS controls for access governance and auditability.
Details
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Cost/unit |
|---|---|
input_tokens_standard | $2.20 |
input_tokens_global_standard | $2.00 |
input_tokens_global_priority | $4.00 |
input_tokens_global_flex | $1.00 |
output_tokens_global_standard | $10.00 |
output_tokens_global_priority | $20.00 |
output_tokens_global_flex | $5.00 |
cache_read_tokens_global_standard | $0.10 |
cache_read_tokens_global_priority | $0.20 |
cache_read_tokens_global_flex | $0.05 |
Dimensions summary
You pay per token used, with no upfront commitment. Charges split by token type: input, output, cache read, and 30-minute cache write. Each type is billed separately. Pricing also varies by processing tier: standard, priority, flex, and ultrafast, where ultrafast runs faster. Global variants route requests across regions. Long-context variants apply when requests use larger context windows. Combining these factors produces each dimension's rate. Cache read and cache write let you reuse context across requests at distinct rates. Your total depends on token volume, the tier you select, and whether you use long-context or global routing.
Top-of-mind questions for buyers
One unit is one token processed by the model. Input tokens cover the text you send; output tokens cover the text the model generates. Both are counted per million tokens and billed separately on each request.
Cache write stores context so later requests can reuse it. Cache read charges apply when a request reuses that stored context. Reused context is billed at the cache read rate instead of the standard input rate, which lowers cost for agents repeating the same context.
Each tier meters the same token types at a different rate tied to processing behavior. Ultrafast generates tokens faster than standard speed. Priority and flex adjust how your requests are handled and queued. You pick a tier per request, and that choice sets which token rates apply.
openai.com
Helpful?
Vendor refund policy
No refunds, all sales are final.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Software as a Service (SaaS)
SaaS delivers cloud-based software applications directly to customers over the internet. You can access these applications through a subscription model. You will pay recurring monthly usage fees through your AWS bill, while AWS handles deployment and infrastructure management, ensuring scalability, reliability, and seamless integration with other AWS services.
Support
Vendor support
For support, visit
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
OpenAI GPT-6 Sol is built to power complex coding and agentic workflows.
OpenAI GPT-6.1 Terra brings intelligence approaching GPT-6 Astra to everyday work on Amazon Bedrock. Build and scale capable agents for coding, computer use, and professional workflows at Terra pricing.
OpenAI GPT-5.6 Sol brings OpenAI flagship GPT-5.6 model for advanced reasoning, coding, scientific research, cybersecurity, and agentic workflows to Amazon Bedrock.
Use Daybreak Blue to access GPT-5.6 Sol for specialized cyber capabilities for verified defenders conducting advanced, authorized cybersecurity work to Amazon Bedrock.
Customer reviews
reviewer2923167
Prompt caching has reduced costs and improved reliability for agentic workflows
Reviewed on Nov 05, 2026
Review from a verified AWS customer
What is our primary use case?
I have used the product for agentic and Openclaw.
How has it helped my organization?
It is an excellent model, one of the most capable with a great cache price, making usage affordable and reliable.
What is most valuable?
The most valuable feature is prompt caching.
What needs improvement?
AWS’s policy of putting these models behind an account manager wall is frustrating.
It’s ridiculous that you can access these models via the provider for free, but with AWS, when you are paying for it, you need to take multiple days to get a callback.
For how long have I used the solution?
I have been using the solution for one month.
Which solution did I use previously and why did I switch?
Previously, I used Claude and Z.AI.
What's my experience with pricing, setup cost, and licensing?
I recommend using prompt caching.
What other advice do I have?
I believe AWS's policy needs to change.
reviewer2920146
Improving text quality has required trade-offs as slow throughput limits real-time production use
Reviewed on Oct 29, 2026
Review from a verified AWS customer
What is our primary use case?
The solution is used in production.
How has it helped my organization?
It didn't improve the organization because speed is a big issue.
What is most valuable?
The solution offers good quality, but the speed is not up to the mark, and the throughput per second is very low.
What needs improvement?
There is room for improvement in the tokens processed per second.
For how long have I used the solution?
I have used the solution for one month.
Which solution did I use previously and why did I switch?
I previously used Opus 5.5 and Sonnet 5.5.
What's my experience with pricing, setup cost, and licensing?
The setup cost is fine.
Which other solutions did I evaluate?
I considered all Claude and OpenAI models as alternate solutions.
What other advice do I have?
It didn't work for me because of latency, but I like the quality cost trade-off.