GPT-6.1 Sol

GPT-6.1 Sol is OpenAI's mid-priced GPT-6 model, released Sep 29, 2026 at $2 per million input tokens and $10 output, which OpenAI says nearly matches GPT-6 Astra on agentic work. In ChatGPT it is only in Work and Codex; the Chat tab uses the earlier GPT-6 Sol.

GPT-6.1 Sol is the model OpenAI recommends for most hard work: released at DevDay on Sep 29, 2026, it is described as "Near-Astra performance for complex work at a lower cost." It costs $2 per million input tokens and $10 per million output tokens, a fifth of GPT-6 Astra, and it is the default model in Codex's config.toml. In ChatGPT it is in Work and Codex, not Chat. Its predecessor, GPT-6 Sol (Sep 22, 2026), is the model behind the Chat tab's GPT-6 on paid plans.

A bright yellow sun with small stars around it on black OpenAI

Quick facts

Released	GPT-6 Sol Sep 22, 2026; GPT-6.1 Sol Sep 29, 2026
API ids	`gpt-6.1-sol`, `gpt-6-sol`
Context window	1,050,000 tokens (922,000 input), 128,000 output; 500,000 by default in Work and Codex
Knowledge cutoff	Apr 30, 2026
Reasoning effort	low, medium (default), high, xhigh, max
API price, GPT-6.1 Sol	$2 input, $0.10 cached, $10 output per 1M tokens, as of Nov 6, 2026
API price, GPT-6 Sol	$2 input, $0.20 cached, $10 output
In ChatGPT	6.1 Sol: Work and Codex. 6 Sol: Chat (Instant to Extra High), Work, Codex

GPT-6 Sol and GPT-6.1 Sol: what is the difference?

GPT-6 Sol arrived with GPT-6 Luna on Sep 22, 2026, at half the token price of GPT-5.6 Sol ($2 against $4 input). A week later OpenAI shipped GPT-6.1 Sol at the same $2 / $10, but with cached input at $0.10 instead of $0.20, "95% below standard input pricing". OpenAI's DevDay figures, as Unite.AI reported them, claim GPT-6.1 Sol cut factual-error responses from 11.4% (GPT-6 Sol) to 7.7% and more than doubled GPT-6 Sol's Terminal-Bench Science score at max effort.

In the Chat tab, the older model still does the work: "GPT-6 Sol powers Instant, Medium, High, and Extra High" when GPT-6 is selected on Plus, Pro, Business and Enterprise. ChatGPT's pricing table lists the two together as "GPT-6.1 Sol / GPT-6 Sol" for Plus and Pro.

Where is GPT-6.1 Sol available?

ChatGPT Work and Codex: Plus, Pro, Business, Enterprise and Edu, rolled out Pro first, then the others, from Sep 29, 2026. On Enterprise and Edu it is off until an administrator enables it. Free and Go do not get it.
Not in Chat: OpenAI's help center is plain that GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Sol and GPT-6 Luna as Work and Codex models "are not available in regular ChatGPT conversations". (GPT-6 Sol reaches Chat only as the engine of "GPT-6".)
OpenAI API: as gpt-6.1-sol, with the Responses API tools, Chat Completions and Batch; no fine-tuning (OpenAI API docs).

How much can you use it?

OpenAI estimates that a Plus or Business Standard user gets 15 to 160 local GPT-6.1 Sol messages per five hours in Work and Codex (15 to 150 on GPT-6 Sol), and that Pro has no five-hour limit. Beyond that, ChatGPT Credits pay for more at 50 credits per million input tokens, 2.5 cached and 250 output. GPT-6.1 Sol Ultrafast came to Codex and Work on Oct 8, 2026 for Pro $500 and eligible Enterprise and Edu; in the API Ultrafast costs $12 / $60 per 1M tokens, and Fast mode $4 / $20.

Benchmarks

OpenAI says GPT-6.1 Sol nearly matches GPT-6 Astra on agentic coding, computer use and professional work at one-fifth of Astra's token prices. Its DevDay figures, relayed by Unite.AI: on OSWorld 2.0 (offline) it came within 2.1 points of Astra at about one-seventh the cost per task, and on AutomationBench it scored 2.2 points above Claude Opus 5.5 at medium effort. These are OpenAI's own claims; the full announcement table was not read.

Independent, dated results:

Artificial Analysis Intelligence Index (v5, read Nov 6, 2026): GPT-6.1 Sol (max) 55, two points under Astra, at $1.27 per index task against Astra's $6.24.
LMArena Text (Oct 27, 2026): 27th, 1481, ten places above Astra.
LMArena WebDev (Oct 27, 2026): 5th, 1751.
LMArena Vision (Oct 19, 2026): 9th, 1294, the best OpenAI model there.

See OpenAI Benchmarks for the rivals around it.

Frequently asked questions

Is GPT-6.1 Sol in ChatGPT? Yes, but only in ChatGPT Work and Codex, on Plus, Pro, Business, Enterprise and Edu. The Chat tab's GPT-6 runs on the earlier GPT-6 Sol.

How much does GPT-6.1 Sol cost in the API? GPT-6.1 Sol costs $2 per million input tokens, $0.10 cached and $10 output for prompts up to 272K tokens, as of Nov 6, 2026. Longer prompts cost $4 / $0.20 / $15, and the Batch API halves the price.

Is GPT-6.1 Sol as good as GPT-6 Astra? Close, by OpenAI's account and by Artificial Analysis's index (55 against 57), for about a fifth of the price per token. Astra stays ahead on the hardest tasks in OpenAI's own charts.

What is the difference between GPT-6 Sol and GPT-6.1 Sol? GPT-6.1 Sol, a week newer, has cheaper cached input and, by OpenAI's figures, fewer factual errors and much better science-task scores. Both cost $2 / $10 per million tokens; GPT-6 Sol still powers GPT-6 in ChatGPT's Chat tab.