What is GPT-6 Sol?
GPT-6 Sol is the middle model in OpenAI's GPT-6 family, between Astra and Luna, built for complex coding and agent workflows. It reads text and images, writes text, and works across a 1.05M-token context with up to 128K tokens of output; its knowledge stops at April 2026. It costs a fifth of GPT-6 Astra per token and keeps the same context, output limit and API tools.
It is a current model and the successor to GPT-5.6 Sol, which stays on sale. OpenAI recommends it for everyday and complex work in Codex and ChatGPT Work on Plus and higher plans, but it is not offered in ChatGPT's Chat.
What changed from GPT-5.6 Sol?
Lower token prices, newer knowledge and new API controls; OpenAI also says its answers are more reliable and clearer.
- Price
- It costs $2 / $10 per million tokens, against GPT-5.6 Sol's $5 / $30.
- Knowledge
- Its knowledge stops at April 2026, against February 2026 for GPT-5.6 Sol.
- API controls
- Mid-turn steering and effort changes that keep the prompt cache are GPT-6 features that GPT-5.6 Sol lacks.
- Where it runs
- It is in Codex and ChatGPT Work but, unlike GPT-5.6 Sol, not in ChatGPT's Chat.
Who is GPT-6 Sol best for?
GPT-6 Sol is best for everyday coding in Codex, API agents that need OpenAI's full tool set at a lower price, planning and review next to GPT-6 Luna, and teams on Amazon Bedrock.
Everyday coding in Codex
OpenAI suggests starting at Sol Medium for everyday and complex code changes, and it comes with the $20-a-month Plus plan.
API agents on a budget
It keeps Astra's full Responses API tool set, including computer use, a hosted shell and MCP, at $2 / $10 per million tokens instead of $10 / $50.
Planning and review next to Luna
Many developers run Sol as the planner or reviewer and hand the well-scoped coding to the much cheaper GPT-6 Luna.
Teams on Amazon Bedrock
It is on Bedrock in the us-east-1 region, and OpenAI says Bedrock's price in commercial regions matches its own.
What's included with GPT-6 Sol?
Six effort levels, pro mode, mid-turn steering, effort changes that keep the cache, a full set of hosted tools, and cheaper Batch and Flex processing.
- Six effort levels
- Reasoning runs from none to max with medium as the default; none also allows function calling in Chat Completions.
- Pro mode
- Setting the reasoning mode to pro makes it do more work on hard tasks, billed at the normal token rates.
- Mid-turn steering
- Over a WebSocket you can send a correction while it works, and the Responses API keeps the work already done.
- Effort changes that keep the cache
- A configuration_update item raises or lowers effort mid-conversation without rewriting the cached prompt.
- Hosted tools
- Web and file search, code interpreter, a hosted shell, apply patch, skills, computer use, MCP and tool search all work on the Responses API.
- Batch, Flex and Fast mode
- Batch and Flex cost half the standard rate for work that can wait; Fast mode costs double.
How much does GPT-6 Sol cost?
GPT-6 Sol costs $2 per million input tokens and $10 per million output tokens on the OpenAI API, and in ChatGPT it comes with Plus at $20 a month and higher paid plans, not with Free or Go.
- Cached input $0.20, cache writes $2.50 per million tokens
- Prompts over 272K tokens: $4 / $15 for the whole request
- Batch and Flex at half price, Fast mode at double
- About 15 to 150 GPT-6 Sol messages every five hours
- In Codex and ChatGPT Work, not in Chat
- Buy ChatGPT credits when you run out
- Same GPT-6 Sol allowance as Plus
- Admin controls, SAML SSO and MFA
- No training on business data by default
- 5x or 20x the usage of Plus
- About 70 to 700, or 300 to 3,000, GPT-6 Sol messages every five hours
On the API, Batch and Flex cost $1 / $5 per million tokens and Fast mode $4 / $20; data-residency endpoints add 10%, and the free API tier does not include GPT-6 Sol. ChatGPT Business is $20 per user per month billed yearly, for 2 or more users; Enterprise and Edu are priced by sales. ChatGPT Free ($0) and Go ($8 a month) get GPT-6 Luna, not Sol. The message counts are OpenAI's estimates for Codex and ChatGPT Work, not fixed limits. Full pricing on developers.openai.com. We checked these prices on .
How do readers rate GPT-6 Sol?
No one has rated GPT-6 Sol yet. Be the first.
Sign in to rate. One rating per account.
GPT-6 Sol pros and cons
It keeps Astra's context and tools at a fifth of the price, but early users are split on whether it beats GPT-5.6 Sol, and some say it spends more reasoning tokens per task.
Pros
- At $2 / $10 per million tokens it costs a fifth of GPT-6 Astra while keeping the same 1.05M-token context, 128K output and API tools.
- Several early users found it more efficient than GPT-5.6 Sol and say it gets coding and agent work done in less time.
- It pairs well with GPT-6 Luna: developers use Sol to plan and review and Luna to write the code, which stretches ChatGPT limits further.
- OpenAI says it gives more reliable facts and clearer answers than GPT-5.6 Sol, and its knowledge is two months newer.
Cons
- Early reviews are split: several Codex users call it a step back from GPT-5.6 Sol on complex work and have switched back.
- It can spend more reasoning tokens and time than GPT-5.6 Sol, so the saving per task is smaller than the per-token price suggests.
- Some users report it fumbling small, clearly defined jobs, such as an Excel formatting change that took seven attempts.
- In ChatGPT it runs only in Codex and ChatGPT Work, not Chat, and a prompt over 272K tokens bills the whole request at $4 / $15.
What are the best GPT-6 Sol alternatives?
Four tools on Recaply do GPT-6 Sol's job. Reflection Beam is open source, so you can run it on your own server.
GPT-6 AstraPaidChoose it when a job needs OpenAI's strongest model across many steps and tools; it costs $10 / $50 per million tokens, five times Sol, with the same 1.05M-token context.
Claude Opus 5.5PaidChoose it when you want Anthropic's default coding model and newer knowledge: it costs $4 / $20 per million tokens for a 1M-token context, and its reliable knowledge runs to June 2026.
Gemini 3.8 FlashFreeChoose it when you want a free API tier to test on, or need video and audio input, which GPT-6 Sol does not take.
Reflection BeamOpen SourceChoose it when you want a 501B open-weight model with 23B active for coding and agents, self-hosted under Apache 2.0.
Questions about GPT-6 Sol
What is GPT-6 Sol?
GPT-6 Sol is OpenAI's mid-tier GPT-6 model for complex coding and agent work, sold on the API at $2 / $10 per million tokens and in paid ChatGPT plans through Codex.
Who is GPT-6 Sol for?
GPT-6 Sol is best for everyday coding in Codex, API agents that need OpenAI's full tool set at a lower price, planning and review next to GPT-6 Luna, and teams on Amazon Bedrock.
Is GPT-6 Sol free?
No. ChatGPT Free and Go do not include it, and the OpenAI API's free tier does not cover it. You need ChatGPT Plus at $20 a month or a higher plan, where it runs in Codex and ChatGPT Work, or paid API usage at $2 / $10 per million tokens.
How much does GPT-6 Sol cost?
GPT-6 Sol's paid plans are API at $2 / $10 per million input / output tokens, ChatGPT Plus at $20 per month, ChatGPT Business at $25 per user per month and ChatGPT Pro at From $100 per month.
Is there an open source alternative to GPT-6 Sol?
Yes. Reflection Beam is an open source tool that does the same job, so you can run your own.
What are the best GPT-6 Sol alternatives?
The closest alternatives on Recaply are GPT-6 Astra, Claude Opus 5.5, Gemini 3.8 Flash and Reflection Beam.
Where can I try GPT-6 Sol?
On its official website, developers.openai.com.
Is GPT-6 Sol better than GPT-5.6 Sol?
OpenAI says GPT-6 Sol gives more reliable facts and clearer answers than GPT-5.6 Sol, and it costs less per token. Early users are split: some find it faster and more efficient, while others went back to GPT-5.6 Sol for complex coding, so test both on your own tasks.
Can I use GPT-6 Sol in ChatGPT?
Yes, on the Plus, Pro, Business, Enterprise and Edu plans, in Codex and ChatGPT Work but not in Chat. Enterprise administrators have to turn it on first. ChatGPT Free and Go get GPT-6 Luna instead, in the desktop app.
GPT-6 Sol vs GPT-6 Astra: which should I use?
Start with GPT-6 Sol for everyday and complex work; OpenAI suggests Astra when a task needs the strongest capability across many steps and tools. Astra costs five times as much per token, $10 / $50 against $2 / $10 per million tokens.
Did GPT-6 Sol have a problem reading images?
Yes, at first. OpenAI fixed a bug in image encoding that degraded image understanding in GPT-6 Sol and GPT-6 Luna, including computer use, and recommends rerunning evaluations and retrying image workflows that were affected.
Our verdict
GPT-6 Sol is a sensible default for OpenAI API work and Codex, but test it against GPT-5.6 Sol before you move complex jobs.
It keeps Astra's 1.05M-token context and full tool set at a fifth of the price, yet early users are split on whether it beats GPT-5.6 Sol, and some say it spends more reasoning tokens per task.
- Skip it for
- Anyone who wants it in ChatGPT's Chat, where it is not offered.
- Check first
- Compare the cost per finished task, not per token, on a few of your real jobs.
GPT-5.5
GPT-5.5-Cyber
Grok 4.7
Step 5 Preview
GPT-5.6