DEV Community

Cover image for Is GPT-6.1 Sol Free?
Hassann
Hassann

Posted on Originally published at apidog.com

Is GPT-6.1 Sol Free?

No, GPT-6.1 Sol isn’t free. OpenAI launched it on September 29, 2026 for Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex. It isn’t in Chat yet, Free and Go plans don’t get it, and the API model gpt-6.1-sol has no free tier. Standard API pricing is $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens.

Try Apidog today

The cheaper routes did get cheaper. GPT-6.1 Sol keeps GPT-6 Sol’s $2/$10 list price and halves the cached-input rate. This guide compares each route, shows what is not free, calculates a real cost example, and points to a free alternative in the GPT-6 family.

For the model itself, see what GPT-6.1 Sol is. For its predecessor, see is GPT-6 Sol free. Before choosing a route, measure your own prompt costs in Apidog.

Every route to GPT-6.1 Sol, compared

Route Cost What you get The catch
Free alternative (not Sol): GPT-6 Luna in the ChatGPT desktop app $0 on the Free plan Luna, not Sol, in Work and Codex in the desktop app Different model; not in Chat; no API key
ChatGPT Plus in Codex or ChatGPT Work A Plus subscription 6.1 Sol for coding and work tasks Plan usage limits; not in Chat; no API key
Sign in with ChatGPT partner apps Counts against your Plus or Pro usage Your plan inside apps such as Devin, Notion, Vercel, T3, OpenClaw, and Dactyl Plus and Pro only; per-app weekly cap; credits off by default
API, Batch, or Flex $1 input, $0.05 cached, $5 output Half the Standard rate Batch is async; Flex is slower and can return 429
API, Standard with caching $2 input, $0.10 cached, $10 output The model from your own code Every token billed; cache hits need a stable prefix
OpenRouter openai/gpt-6.1-sol $2 input, $0.10 cached, $10 output (OpenAI endpoint) One key across providers Same list price; not free

All API prices are per million tokens for prompts up to 272K input tokens, based on OpenAI’s pricing page and OpenRouter’s model page.

ChatGPT Plus: the entry plan for interactive use

If you want to use 6.1 Sol yourself in Codex or ChatGPT Work, Plus is the lowest plan that includes it.

OpenAI’s launch post makes it available “to all Plus, Pro, Business, Enterprise, and Edu users in ChatGPT Work and Codex” and says it “is not yet available in Chat.” According to OpenAI’s models docs, Enterprise and Edu users need an admin to enable it.

You do not pay per token in ChatGPT, but plan usage limits still apply. See Codex usage limits for how those limits work.

The free Codex guide covers no-cost Codex options, but GPT-6.1 Sol starts at Plus.

Sign in with ChatGPT: use your plan inside other apps

DevDay introduced Sign in with ChatGPT. The identity layer, based on OAuth/OIDC, is available globally. Plus and Pro users can also use their plan allowance in participating tools.

OpenAI launched with 16 partners. Its DevDay recap names Cognition’s Devin, Notion, Vercel, T3, OpenClaw, and Dactyl.

According to OpenAI’s help article on using your plan in other apps:

  • Eligible requests count toward your ChatGPT Work and Codex usage.
  • Each app has a weekly cap expressed as a percentage of your total weekly usage.
  • The cap is not a reserved pool of usage.
  • Paying with credits after hitting the cap is optional, off by default, and available only after the app limit reaches 100%.
  • The app receives your name, email address, and profile picture. It does not receive conversations, memories, or an API key.

Check which model an app uses before assuming it runs GPT-6.1 Sol. For OpenClaw pricing, see is OpenClaw free. For implementation details, including client IDs and OAuth, see the Sign in with ChatGPT guide.

A ChatGPT plan is not an API key

This is the common trap in “free GPT” searches: a Plus or Pro plan gives you GPT-6.1 Sol in ChatGPT Work and Codex, not an API key.

A script, CI job, or agent sending:

Authorization: Bearer $OPENAI_API_KEY
Enter fullscreen mode Exit fullscreen mode

to:

https://api.openai.com/v1
Enter fullscreen mode Exit fullscreen mode

pays per token regardless of your ChatGPT plan.

The exception is Sign in with ChatGPT’s OAuth flow. OpenAI’s plan usage docs make this available to open-source and locally hosted apps, with preview constraints:

  • Streaming only
  • store: false
  • No temperature
  • No hosted tools such as file search

These requests still count against the user’s Plus or Pro usage and the app-specific limit.

The cheapest API route: Batch, Flex, and caching

Use three levers to reduce token costs:

  • Batch runs jobs asynchronously at half the Standard rate: $1 input, $0.05 cached input, and $5 output per million tokens. See the Batch API guide for an end-to-end workflow.
  • Flex uses Batch rates for ordinary requests when you set service_tier: "flex". The Flex processing guide warns that responses can be slower and may return 429 Resource Unavailable errors. Those errors are not charged.
  • Prompt caching is enabled by default. Cached input costs $0.10 per million tokens: 95% below Standard input and half the GPT-6 Sol cached-input rate. OpenAI’s prompt caching guide specifies a 1,024-token minimum, cache writes at 1.25x the uncached input rate ($2.50 per million here), and a reusable prefix lifetime of 30 minutes after the last write or reuse.

Keep requests under 272K input tokens. Above that threshold, the model page says the entire request bills at 2x input and cache rates and 1.5x output rates.

Worked example: the cost of one agent call

Assume an agent step has:

  • A 100,000-token stable prefix: system prompt, tool schemas, and repository context
  • 8,000 new input tokens
  • 2,000 output tokens, including reasoning
  • A cache hit for the stable prefix
  • Explicit-only caching mode, so new tokens are billed as standard input

Standard pricing with a cache hit

  • Cached input: 100,000 × $0.10 / 1,000,000 = $0.0100
  • New input: 8,000 × $2.00 / 1,000,000 = $0.0160
  • Output: 2,000 × $10.00 / 1,000,000 = $0.0200
  • Total: $0.0460 per call

Batch or Flex pricing

Every rate is halved:

$0.0050 + $0.0080 + $0.0100 = $0.0230 per call
Enter fullscreen mode Exit fullscreen mode

GPT-6 Sol comparison

On GPT-6 Sol, the cached-input line doubles to $0.0200:

$0.0200 + $0.0160 + $0.0200 = $0.0560 per call
Enter fullscreen mode Exit fullscreen mode

Without a cache hit on GPT-6.1 Sol, all 108,000 input tokens are standard input:

108,000 × $2.00 / 1,000,000 = $0.2160
$0.2160 + $0.0200 output = $0.2360 per call
Enter fullscreen mode Exit fullscreen mode
Scenario Per call 1,000 calls
6.1 Sol Standard, cache hit $0.046 $46
6.1 Sol Batch or Flex, cache hit $0.023 $23
GPT-6 Sol Standard, cache hit $0.056 $56
6.1 Sol Standard, no cache hit $0.236 $236

The first request writes the prefix to cache at $2.50 per million tokens:

100,000 × $2.50 / 1,000,000 = $0.25
Enter fullscreen mode Exit fullscreen mode

After that first write:

  • Cache hits reduce the input bill by roughly fivefold.
  • Batch or Flex halves the total rate.
  • Moving from GPT-6 Sol saves $0.01 per call in this example because only cached-input pricing changed.

For cache implementation guidance, read GPT-6 prompt caching.

What is not free

  • ChatGPT Free and Go: no GPT-6.1 Sol.
  • ChatGPT Chat: GPT-6.1 Sol is not available there yet on any plan.
  • The API: no free tier.
  • Sign in with ChatGPT on a Free account: plan usage is available to Plus and Pro only.
  • OpenRouter: same list price, not a discount.
  • Ultrafast: GPT-6.1 Sol’s option is “coming soon.” On GPT-6 Astra, it costs 6x Standard.

The free alternative: GPT-6 Luna

If you need a no-cost option, GPT-6 Luna is available on the Free plan in Work and Codex through the ChatGPT desktop app. Go, a paid plan, includes it too.

Luna is not Sol. It is not available in Chat, and a plan seat does not include an API key, so it cannot directly power your script or agent.

See the free GPT-6 Luna guide for setup.

On the API, Luna lists at $0.10 input and $0.50 output per million tokens: one-twentieth of GPT-6.1 Sol’s input and output rates. That makes Luna the lower-cost choice for high-volume classification and extraction.

Measure real token usage in Apidog before you choose

The example above uses hypothetical token counts. Your actual prompt sizes, cache hit rate, and output volume determine the cheapest route.

Use a saved request in Apidog to measure them:

  1. Create an environment with OPENAI_API_KEY and MODEL_ID, setting MODEL_ID to gpt-6.1-sol.
  2. Save a POST request to https://api.openai.com/v1/responses.
  3. Set the Authorization header to Bearer {{OPENAI_API_KEY}}.
  4. Add a body containing "model": "{{MODEL_ID}}", "reasoning": {"effort": "low"}, and a real workload prompt as input.
  5. Send the request twice.
  6. Assert that $.usage.input_tokens_details.cached_tokens is greater than 0 on the second run. This makes a prompt edit that breaks caching fail the test.
  7. Compare usage.output_tokens and usage.output_tokens_details.reasoning_tokens at low and medium reasoning effort. The reasoning guide states that reasoning tokens are billed as output.
  8. Add a post-response script to calculate Standard pricing: 
const u = pm.response.json().usage;
const d = u.input_tokens_details || {};
const cached = d.cached_tokens || 0;
const written = d.cache_write_tokens || 0;
const fresh = u.input_tokens - cached - written;

const usd =
  (fresh * 2 + cached * 0.1 + written * 2.5 + u.output_tokens * 10) / 1e6;

console.log(`gpt-6.1-sol Standard: $${usd.toFixed(4)} per call`);
Enter fullscreen mode Exit fullscreen mode

Multiply the result by daily volume, then halve it for Batch or Flex.

Save the request in a test scenario and rerun it through the Apidog CLI in CI. If a prompt change drops your cache hit rate, you can detect it before it affects your invoice.

FAQ

Is GPT-6.1 Sol free in ChatGPT?

No. It is available in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu users. It is not in Chat yet and is not available on Free or Go.

Can I use my ChatGPT Plus plan to call the GPT-6.1 Sol API?

Not with an API key. A ChatGPT plan is not an API key, and API calls pay per token. Sign in with ChatGPT lets eligible apps use an OAuth flow so requests can count against Plus or Pro usage, subject to preview limits.

Is there a free GPT-6.1 Sol API?

No. The cheapest official option is Batch or Flex at $1 input, $0.05 cached input, and $5 output per million tokens.

Is GPT-6.1 Sol free on OpenRouter?

No. OpenRouter lists openai/gpt-6.1-sol at $2/$10 on its OpenAI endpoint, matching OpenAI’s list price.

Is GPT-6.1 Sol cheaper than GPT-6 Sol?

The list price is the same: $2 input and $10 output per million tokens. Cached input is half the price: $0.10 versus $0.20, so high-cache-hit workloads cost less.

Next step

Choose the route using your own usage data. Download Apidog, save one real request, send it twice, and inspect usage.

  • For interactive GPT-6.1 Sol use, Plus in Codex is the entry point.
  • For code, use the API with prompt caching.
  • Move work that can wait to Batch or Flex.

For the first API request and migration from gpt-6-sol, see the GPT-6.1 Sol API guide.

Top comments (0)