GPT-6.2 Sol
- Provider
- OpenAI
- Status
- Current
- Context
- 1,050,000 tok
- Price
- $2 / $10 /MTok
- Knowledge
- 2026-06-01
GPT-6.2 Sol is the model OpenAI describes as efficient and versatile for everyday work, released on 17 November 2026 alongside GPT-6.2 Astra (OpenAI). It succeeds GPT-6.1 Sol at the same $2/$10 per million tokens, a quarter of GPT-6.2 Astra’s price.
Artificial Analysis scored it 57 on the Intelligence Index v5.0.1 at max effort at launch, two points above GPT-6.1 Sol and level with GPT-6 Astra, at $1.01 per completed index task against GPT-6.1 Sol’s $1.27 and Astra’s $6.24.
Quick specs
| GPT-6.2 Sol | |
|---|---|
| Released | 17 November 2026 |
| Input price | $2.00 / MTok |
| Cached input | $0.10 / MTok |
| Output price | $10.00 / MTok |
| Over 272K input tokens | $4.00 in / $15.00 out |
| Context window | 1,050,000 tokens (922,000 input) |
| Max output | 128,000 tokens |
| Knowledge cutoff | 1 June 2026 |
| Input / output | Text and images in, text out |
| AA Intelligence Index | 57 (max) at launch (v5.0.1), 55 (high), 53 (medium) |
| Cost per AA index task | $1.01 |
| API model ID | gpt-6.2-sol |
Specs and prices are from OpenAI’s model page and API pricing.
What changed from GPT-6.1 Sol
Intelligence. At launch, up two points on Artificial Analysis’s index v5.0.1, from 55 to 57 at max effort, level with GPT-6 Astra.
Cost per task. Down about a fifth on Artificial Analysis’s index, from $1.27 to $1.01 per completed task, at the same token price.
Ultrafast. Pro 500 subscribers can use OpenAI’s Ultrafast mode with both GPT-6.2 models, and OpenAI says it is coming to the $200 Pro plan, starting with GPT-6.2 Sol (OpenAI).
Knowledge. Its knowledge cutoff moves to 1 June 2026, against 30 April for GPT-6.1 Sol.
Benchmark performance
| At GPT-6.2 Sol’s launch (17 Nov 2026) | Intelligence Index v5.0.1 (max) | Cost per index task | Price (in/out per MTok) |
|---|---|---|---|
| GPT-6.2 Sol | 57 | $1.01 | $2 / $10 |
| GPT-6 Astra | 57 | $6.24 | $10 / $50 |
| GPT-6.1 Sol | 55 | $1.27 | $2 / $10 |
| Claude Sonnet 5.5 | 55 | $10.02 | $2 / $10 |
| Gemini 4 Hydrogen (xhigh) | 54 | $3.16 | $1.50 / $7.50 |
Source: Artificial Analysis, independent; GPT-6 Astra and GPT-6.1 Sol from their model pages (GPT-6 Astra, GPT-6.1 Sol).
OpenAI’s own figures, vendor-run: it says that on FrontierCode 2.0 GPT-6.2 Sol nearly matches GPT-6.2 Astra on merge-ready changes to real codebases, at about a fifth of the cost (OpenAI), and its system card puts it at 96.0% on GPQA Diamond without tools. OpenAI’s launch post gives no SWE-bench figure.
Safety testing
OpenAI’s system card rates GPT-6.2 Sol Critical for cybersecurity and High for biological and chemical capabilities, like GPT-6.2 Astra. Unlike Astra, it falls below the High threshold for AI self-improvement. Both models were better than GPT-6 Astra and GPT-6.1 Sol at turning known software vulnerabilities into working exploits, and both took improper actions about half as often as the models they replace in simulated workplace tasks. Both were also more likely to leave known obstacles out when reporting their work, which OpenAI says the Codex developer prompt substantially reduces. GPT-6.2 Sol showed fewer signs of knowing it was being evaluated than GPT-6 Astra and GPT-6.1 Sol.
How it compares
vs GPT-6.2 Astra
GPT-6.2 Astra scored five points higher at launch (62) at four times the token price ($8/$40), and a completed index task cost $4.00 against Sol’s $1.01.
vs Claude Sonnet 5.5
Claude Sonnet 5.5 has the same $2/$10 token price. When GPT-6.2 Sol launched, Sonnet 5.5 scored 55 on the index, two points lower, and a completed max-effort task cost $10.02, about ten times Sol’s.
vs Gemini 4 Hydrogen
Gemini 4 Hydrogen has lower token prices ($1.50/$7.50). When GPT-6.2 Sol launched, Hydrogen scored 54 at xhigh, three points lower, and a completed task cost $3.16, three times Sol’s.
Known limitations
No SWE-bench figure. OpenAI’s launch post gives no SWE-bench Pro or Verified result.
Not in regular chat. At launch it was in ChatGPT Work and Codex, not ordinary ChatGPT conversations.
Cyber help is restricted. OpenAI rates it Critical for cybersecurity and launched it with tighter safeguards on higher-risk security requests.
Frequently asked questions
When was GPT-6.2 Sol released?
17 November 2026, on the API as gpt-6.2-sol and in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users. At launch it was not in regular ChatGPT chat.
How much does GPT-6.2 Sol cost?
$2 per million input tokens and $10 per million output tokens, the same as GPT-6.1 Sol and a quarter of GPT-6.2 Astra. Cached input is $0.10 per million. Requests over 272K input tokens cost $4 on input and $15 on output.
Is GPT-6.2 Sol better than GPT-6.1 Sol?
On Artificial Analysis’s index, yes, at the same price. At launch it scored 57 against GPT-6.1 Sol’s 55 on the Intelligence Index v5.0.1, with a completed task costing $1.01 against $1.27.
Is GPT-6.2 Sol as good as GPT-6 Astra?
On Artificial Analysis’s index it scored the same at launch, 57, at about a sixth of the cost per completed task ($1.01 against $6.24) and a fifth of the token price. For current standings, see best AI models.