GPT-6.1 Sol: lower prices for complex coding in iolys
GPT-6.1 Sol halves cached input prices versus GPT-6 Sol and costs less than Astra. Compare API prices and try it on C# and .NET projects in iolys.
GPT-6.1 Sol brings a new option for complex coding at a lower cost than Astra. OpenAI released it on September 29, 2026, positioning it close to Astra in capability. For developers who already use GPT-6 Sol, the clearest price change is cached input is 50% cheaper. OpenAI release details.
Here is what changed, how the savings work, and how to try the model on your C# and .NET projects through Codex in iolys for Visual Studio 2026.
What GPT-6.1 Sol brings to coding
OpenAI describes GPT-6.1 Sol as delivering near-Astra performance for complex coding, computer use, and professional work. That is OpenAI's assessment; compare the models on your own tasks to see whether Sol meets your needs. GPT-6.1 Sol model details.
Try it on your C# and .NET projects to fix bugs, add features, or refactor code.
The model retains a 1.05-million-token context window and a maximum output of 128,000 tokens. These limits match GPT-6 Sol, giving you room for substantial code and conversation context. GPT-6.1 Sol specifications, GPT-6 Sol specifications.
A much lower price than Astra
The following prices are in USD per million text tokens, for OpenAI's direct API at Standard speed with prompts of up to 272,000 input tokens. Checked on October 2, 2026.
| Token category | GPT-6 Astra | GPT-6.1 Sol | Change vs. Astra |
|---|---|---|---|
| Input | $10.00 | $2.00 | ↓ −80% |
| Cached input (reads) | $1.00 | $0.10 | ↓ −90% |
| Cache writes | $12.50 | $2.50 | ↓ −80% |
| Output | $50.00 | $10.00 | ↓ −80% |
GPT-6.1 Sol offers 80% lower input and output prices and 90% lower cache-read prices than Astra. GPT-6 Astra pricing, GPT-6.1 Sol pricing.
This makes Sol worth evaluating for repeated complex tasks where Astra's price is harder to justify. The cost of completing a task still depends on how many tokens and attempts each model needs.
Conditions that change the API bill
For GPT-6.1 Sol, prompts above 272,000 input tokens use higher rates for the full request: $4 input, $0.20 cached input, $5 cache writes, and $15 output per million tokens. Fast mode costs twice Standard; Batch and Flex cost half Standard. Regional processing adds 10% where available, and paid tools can add charges. OpenAI API pricing.
Try GPT-6.1 Sol in Visual Studio with iolys
iolys connects the Codex CLI to your active Visual Studio 2026 solution. You can use eligible ChatGPT subscription access while keeping files, tool activity, and permissions in the IDE. Codex in iolys.
- Open Manage Providers in iolys and select Codex.
- Install or update the locally managed CLI, then connect your ChatGPT account.
- Choose GPT-6.1 Sol when it appears in your model options and try a concrete C# or .NET task.
OpenAI's launch rollout includes Plus, Pro, Business, Enterprise, and Edu. Availability depends on your account, client, and workspace settings; Enterprise and Edu administrators must enable the model. Official model availability.
To evaluate GPT-6.1 Sol on your own solution, Install iolys for Visual Studio.
Share your experience on Discord
Tried GPT-6.1 Sol on a C# or .NET project? Join us on Discord to share your results, ask about setup, or suggest improvements to iolys.