OpenAI

GPT-6.2 Astra

OpenAI's higher-capability general model, released 2026-11-17 under API id gpt-6.2-astra, at a Standard short-context API price of $8 input and $40 output per million tokens — 20% below GPT-6 Astra, with a 60% discount on cached input, per OpenAI — and a 1,050,000-token context. Its knowledge cutoff is May 18, 2026, earlier than GPT-6.2 Sol's June 1, 2026.

GPT-6.2 Astra benchmarks and pricing, every number sourced: no GPT-6.2 Astra benchmark scores tracked yet, priced at $8.00 per million input tokens and $40.00 per million output.

Released
2026-11-17
Availability
API id gpt-6.2-astra. ChatGPT Work and Codex for GPT-6.2 Astra cover Plus, Pro, Enterprise, and Edu, and eligible Enterprise and Edu users get that product access once a workspace admin enables it. ChatGPT Ultrafast on Pro 500 includes GPT-6.2 Astra; Pro 200 ChatGPT Ultrafast starts with GPT-6.2 Sol only. API Ultrafast is available for gpt-6.2-astra at launch. Checked 2026-11-18.
License
proprietary
Context window
1M tokensvendor-stated (developers.openai.com/api/docs/models/gpt-6.2-astra: 1,050,000-token context, 922,000 max input, 128,000 max output; checked 2026-11-18)
Knowledge cutoff
2026-05-18
Succeeded
GPT-6 Astra (2026-09-03)
Parameters
Not disclosed
Architecture
Not disclosed

GPT-6.2 Astra announced pricing

$8.00 in / $40.00 out per 1M tokens — official pricing source

What GPT-6.2 Astra costs per job

GPT-6.2 Astra cost for three reference workloads, computed from its list rates
WorkloadTokens in / outCost
One long chat turn100K / 10K$1.20
A codebase review1,000K / 100K$12.00
A day of agent work10,000K / 1,000K$120.00

Computed from GPT-6.2 Astra’s list rates above — cache discounts and batch tiers are not applied.

GPT-6.2 Astra is one of 10 OpenAI models tracked on this site, at these official list prices.

OpenAI model pricing, official list rates
ModelIn / 1MOut / 1M
GPT-6.2 Sol$2.00$10.00
GPT-6.2 Astra$8.00$40.00
GPT-6 Cyber$20.00$100.00
GPT-6.1 Terra$1.00$5.00
GPT-6.1 Sol(previous version)$2.00$10.00
GPT-6 Luna$0.10$0.50
GPT-6 Sol(previous version)$2.00$10.00
GPT-6 Astra(previous version)$10.00$50.00
GPT-5.6 Luna(previous version)$0.20$1.20
GPT-5.6 Sol(previous version)$4.00$20.00

Notes on the record

OpenAI released GPT-6.2 Astra on 2026-11-17 under API id gpt-6.2-astra. The stored price is Standard short context only: $8 input and $40 output per million tokens, cached input $0.40, cache writes $10 for 30 minutes or $20 for 12 hours. OpenAI calls those input and output rates 20% below GPT-6 Astra, with a 60% discount on cached input. Above 272,000 input tokens the whole GPT-6.2 Astra request is 2x input and cache and 1.5x output: $16, $0.80 cached, $20 or $40 cache write, and $60 output per million tokens. For GPT-6.2 Astra, Batch is 50% of that Standard rate, Fast is 2x at the same context length, and Ultrafast is 3x through 2027-02-28 Pacific time then 6x from 2027-03-01. Those GPT-6.2 Astra mode rates are not stored in price_in or price_out.

GPT-6.2 Astra's window is 1,050,000 tokens, at most 922,000 input and 128,000 output, with text and image in and text out. Its knowledge cutoff is May 18, 2026. GPT-6.2 Astra supports reasoning tokens and rejects reasoning effort none. Chat Completions, Responses, and Batch are supported for gpt-6.2-astra, and fine-tuning is not. OpenAI publishes no parameter count for GPT-6.2 Astra.

ChatGPT Work and Codex offer GPT-6.2 Astra to Plus, Pro, Enterprise, and Edu, and eligible Enterprise and Edu users get that product once a workspace admin enables it. Pro 500 ChatGPT Ultrafast includes Astra. Pro 200 ChatGPT Ultrafast starts with GPT-6.2 Sol only. API Ultrafast is on for gpt-6.2-astra at launch. Admin enablement of ChatGPT Work and Codex is not a stated Enterprise or Edu ChatGPT Ultrafast grant for GPT-6.2 Astra.

No GPT-6.2 Astra score row is stored. The launch post's vendor benchmark claims were not fetched from a tracked board. The Nov 17 changelog excerpt has no deprecation of GPT-6 Astra, GPT-6 Sol, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, or GPT-6 Cyber, and this GPT-6.2 Astra entry does not rewrite them.

Compare with

FAQ

What is GPT-6.2 Astra?

GPT-6.2 Astra is OpenAI's higher-capability general model, released on 2026-11-17 under API id gpt-6.2-astra. GPT-6.2 Astra takes text and image input and returns text. The GPT-6.2 Astra context window is 1,050,000 tokens, with a maximum of 922,000 input tokens and 128,000 output tokens. GPT-6.2 Astra supports reasoning tokens, and requests that set reasoning effort to none are rejected. Chat Completions, Responses, and Batch work for gpt-6.2-astra, and fine-tuning does not. OpenAI has published no parameter count for GPT-6.2 Astra.

How much does GPT-6.2 Astra cost?

GPT-6.2 Astra's stored price is the Standard short-context rate, $8 per million input tokens and $40 per million output tokens. Cached input is $0.40, and cache writes are $10 for 30 minutes or $20 for 12 hours. OpenAI describes that price as 20% below GPT-6 Astra, with a 60% discount on cached input. Above 272,000 input tokens the whole GPT-6.2 Astra request bills at 2x input and cache and 1.5x output. For GPT-6.2 Astra, Batch is half of Standard, Fast is double, and Ultrafast is 3x through 2027-02-28 Pacific time then 6x. Only the $8 and $40 Standard short-context rates are stored.

Who can use GPT-6.2 Astra, including Ultrafast?

In ChatGPT Work and Codex, GPT-6.2 Astra is available to Plus, Pro, Enterprise, and Edu. Eligible Enterprise and Edu users get GPT-6.2 Astra in those products once a workspace admin enables them. Pro 500 can use ChatGPT Ultrafast with either GPT-6.2 model. Pro 200 ChatGPT Ultrafast starts with GPT-6.2 Sol and is not stated for Astra. API Ultrafast is available for gpt-6.2-astra at launch. Enabling the models for Enterprise or Edu ChatGPT does not, on the checked blog, confirm ChatGPT Ultrafast on those plans.

What are GPT-6.2 Astra's context window and knowledge cutoff?

OpenAI's GPT-6.2 Astra model page states a 1,050,000-token context window and a May 18, 2026 knowledge cutoff. GPT-6.2 Astra's maximum input is 922,000 tokens and its maximum output is 128,000. That May 18 cutoff is earlier than GPT-6.2 Sol's June 1, 2026 cutoff. The GPT-6.2 Astra model page was checked on 2026-11-18.

Does this page include GPT-6.2 Astra benchmark scores?

No GPT-6.2 Astra score row is stored. The launch post and blog state vendor benchmark claims for GPT-6.2 Astra, and this scan did not independently fetch a tracked-board page, so those claims are not rows. The GPT-6.2 Astra model page and pricing page were checked on 2026-11-18.

Does GPT-6.2 Astra replace earlier GPT-6 models?

The fetched Nov 17 changelog excerpt and launch materials do not announce a deprecation of GPT-6 Astra, GPT-6 Sol, GPT-6.1 Sol, GPT-6.1 Terra, GPT-6 Luna, or GPT-6 Cyber. This GPT-6.2 Astra entry does not rewrite those records and does not add a deprecation event. The catalog lineup rule does label GPT-6 Astra a previous version of GPT-6.2 Astra, because the names share the Astra tier and the newer date is later. For GPT-6.2 Astra, that label is the lineup rule, not an OpenAI retirement notice.

Further reading