GPT-6.2 Sol is OpenAI's new model, released on 17 November 2026 together with the flagship GPT-6.2 Astra. OpenAI calls it the go-to model for everyday work, from coding to agents running at scale. Its OpenAI price is unchanged from GPT-6.1 Sol, and on Kumo it costs 45% less than the official rate. Here is what changed, what independent measurements show and how to connect it in a couple of minutes.
Context
1.05M tokens
up to 128K tokens in one answer
Price on Kumo
$1.1 / $5.5
input / output per 1M tokens, 45% below OpenAI
Input
text and images
answers in text
Released
17 Nov 2026
on Kumo since 18 November
What GPT-6.2 Sol is
OpenAI splits its line-up into classes: Astra is the strongest, Sol is the main class for everyday work, Terra and Luna are faster and cheaper. GPT-6.2 Sol replaces GPT-6.1 Sol and, since 17 November, opens new chats in ChatGPT Work.
OpenAI promises better results with fewer output tokens, cleaner code and stronger frontends — websites, games, 3D — plus more precise computer use. On a ChatGPT plan, OpenAI says, GPT-6.2 Sol fits more tasks than GPT-6.1 Sol.
The knowledge cutoff is 1 June 2026. Developers call it over the API as gpt-6.2-sol; reasoning effort is adjustable, and the model does not accept effort none.
GPT-6.2 Sol vs GPT-6.1 Sol and GPT-6.2 Astra
| Spec | GPT-6.1 Sol | GPT-6.2 Sol | GPT-6.2 Astra |
|---|---|---|---|
| Class | main | main | flagship |
| Context | 1.05M tokens | 1.05M tokens | 1.05M tokens |
| OpenAI price, input / output | $2 / $10 | $2 / $10 | $8 / $40 |
| Cache read, OpenAI | $0.1 | $0.1 | $0.4 |
| Kumo price, input / output | $1.1 / $5.5 | $1.1 / $5.5 | $4.8 / $24 |
What the measurements show
OpenAI's main case for Sol is price: on FrontierCode 2.0 it lands within half a point of Astra at about a fifth of the cost (OpenAI figures).
Independent Artificial Analysis puts GPT-6.2 Sol at 57 on its Intelligence Index — above the previous flagship GPT-6 Astra (56.9) — at about $1 per index task. Reasoning effort matters a lot: 57 at max, 53 at medium, 41 at low. Vals AI measures 60.6% on Terminal-Bench 4.0.
There is a weak spot too: on the Artificial Analysis Coding Agent Index Sol 6.2 scores 3 points below GPT-6.1 Sol, driven by a drop on one code question-answering test. If your workload is long agentic work on a repository, compare both models on your own tasks — on Kumo they share one key.
What GPT-6.2 Sol costs
| Where | Input | Output | Cache read |
|---|---|---|---|
| OpenAI | $2 | $10 | $0.1 |
| Kumo, paid from balance | $1.1 | $5.5 | $0.06 ($0.055) |
On Kumo GPT-6.2 Sol is 45% cheaper than at OpenAI itself, and in token packages it is charged with a ×1.2 multiplier — the same as GPT-6.1 Sol. Re-reading the same context — a repository, a conversation, a document — costs a twentieth of regular input, so in agent work the final bill is noticeably below the list price.
GPT-6.2 Sol in the Kumo catalogue
The model card with its current price is below; it leads straight to setup.
What it is good for
- Everyday development: code edits, reviews, tests, frontend — where Astra is overkill on price.
- Agents at scale: multi-step chains with tools where the cost per task matters.
- Computer use and screenshots: the model understands interfaces and documents with images.
- Chatbots and text processing: at a quarter of Astra's price, Sol pays off for high request volume.
What Kumo has verified
Before launch Kumo runs probe requests against every model. For GPT-6.2 Sol, plain answers, streaming and tool calling are confirmed over both OpenAI protocols — Chat Completions and Responses. The model then passed live requests through the api.kumorouter.com gateway, so it also works in Responses API clients such as Codex CLI.
How to connect GPT-6.2 Sol
- Sign up in the Kumo console and top up your balance.
- Create an API key and set a spending limit on it.
- Point any OpenAI-compatible client or SDK at the gateway https://api.kumorouter.com/v1.
- Set the model to gpt-6.2-sol and send your first request.
curl -s https://api.kumorouter.com/v1/chat/completions \
-H "Authorization: Bearer $KUMO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "gpt-6.2-sol",
"messages": [
{"role": "user", "content": "Find the bug in this function and suggest a fix: ..."}
],
"max_tokens": 4096
}'
Python example
The official OpenAI SDK works as is — just change the base URL and the key. The example uses the Responses API; Chat Completions works too.
from openai import OpenAI
client = OpenAI(
base_url="https://api.kumorouter.com/v1",
api_key="KUMO_API_KEY",
)
response = client.responses.create(
model="gpt-6.2-sol",
input="Write tests for this module: ...",
)
print(response.output_text)
FAQ
- How does GPT-6.2 Sol differ from GPT-6.1 Sol?
- Per OpenAI, better results with fewer output tokens, cleaner code and more precise computer use. On the independent Artificial Analysis Intelligence Index Sol 6.2 scores 57, but on agentic coding it trails 6.1 Sol by 3 points. The OpenAI price is the same.
- How does GPT-6.2 Sol differ from GPT-6.2 Astra?
- Astra is the flagship for the hardest tasks: 62 against 57 on the Artificial Analysis index. Sol is a quarter of the price and close to Astra on many tasks, so it pays off for everyday work.
- Why is Kumo cheaper than OpenAI?
- That is Kumo's price for GPT-6.2 Sol: 45% below OpenAI's list price — $1.1 against $2 per million input tokens and $5.5 against $10 for output. You pay from your Kumo balance, with no separate OpenAI account.
- Does GPT-6.2 Sol work in Russia?
- Through Kumo, yes: the gateway works without a VPN, and the balance can be topped up with a Russian card, SBP or crypto.
Sources
- OpenAI docs: GPT-6.2 Sol and GPT-6.2 Astra, ChatGPT and Codex changelog, 17 November 2026.
- Launch coverage and OpenAI figures: The Decoder.
- Independent measurements: Artificial Analysis and Vals AI.
- Kumo prices and specs: models page.