GPT-6.2 Astra and Sol launch in the API, Astra 20% cheaper
OpenAI's new frontier and everyday models are in the API, with breaking parameter changes to check first.
OpenAI released GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026, in ChatGPT, Codex and the OpenAI API. Astra is OpenAI's new frontier model for the hardest work, and Sol is its everyday model for coding and agents. Astra's input and output prices are 20% below GPT-6 Astra's, and Sol keeps GPT-6.1 Sol's prices.
What changed
- Model IDs are
gpt-6.2-astraandgpt-6.2-sol, on the Responses API and Chat Completions. Tool calling requires the Responses API. - Both take text and image input and return text, with a 1,050,000-token context window (922,000 max input) and 128,000 max output tokens. Knowledge cutoffs are May 18, 2026 (Astra) and June 1, 2026 (Sol).
- The API rejects reasoning effort
none. Chat Completions requests must droptemperature,top_p,logprobsandtop_logprobs; Responses requests droptemperature,top_pandtop_logprobs. - From GPT-6.2, the API drops reasoning items when it detects edits to earlier conversation history, a model downgrade, or a change of end user or organization. OpenAI recommends keeping the prompt prefix unchanged between turns.
- Both support US and EU data residency with Standard, Fast and Batch processing; Ultrafast is US only. Microsoft made both models generally available in Microsoft Foundry the same day.
OpenAI API pricing
Standard rates per million tokens for prompts up to 272K input tokens, from OpenAI's pricing page on November 17, 2026:
| Model | Input | Cached input | Cache write (30 min) | Output |
|---|---|---|---|---|
gpt-6.2-astra | $8.00 | $0.40 | $10.00 | $40.00 |
gpt-6.2-sol | $2.00 | $0.10 | $2.50 | $10.00 |
gpt-6-astra | $10.00 | $1.00 | $12.50 | $50.00 |
gpt-6.1-sol | $2.00 | $0.10 | $2.50 | $10.00 |
- Over 272K input tokens, input and cache rates double and output costs 1.5 times the short-context rate for the whole request.
- Batch costs 50% of Standard and Fast mode costs twice Standard.
- Ultrafast costs 3 times Standard through February 28, 2027, and 6 times Standard from March 1, 2027.
- A 12-hour cache write costs $20 (Astra) or $5 (Sol) per million tokens.
What to do
Swap the model ID, then follow OpenAI's migration checklist:
gpt-6-astra -> gpt-6.2-astra
gpt-6.1-sol -> gpt-6.2-sol
reasoning effort none or minimal -> low
remove temperature, top_p, logprobs, top_logprobs
send tool calls through the Responses APIOn RouterPlus
Update, November 18, 2026: GPT-6.2 Astra and Sol are now in the RouterPlus catalog as gpt-6.2-astra and gpt-6.2-sol, served through OpenAI's direct API with a 1,050,000-token context window and 128,000 max output tokens. AWS also made both models generally available on Amazon Bedrock on November 17. Per million tokens from the RouterPlus catalog on November 18, 2026:
| Model | Provider | Input | Cached input | Output |
|---|---|---|---|---|
gpt-6.2-astra | OpenAI (direct API) | $8.00 | $0.40 | $40.00 |
gpt-6.2-sol | OpenAI (direct API) | $2.00 | $0.10 | $10.00 |
RouterPlus bring your own key works with your own OpenAI key, with no RouterPlus fee on those calls. See all OpenAI models on RouterPlus.
Sources
- OpenAI: Introducing GPT-6.2 Astra and Sol (November 17, 2026)
- OpenAI API docs: GPT-6.2 Astra
- OpenAI API docs: GPT-6.2 Sol
- OpenAI API docs: Using GPT-6.2 (migration quickstart)
- OpenAI API docs: Pricing
- Microsoft Foundry Blog: GPT-6.2 Astra and GPT-6.2 Sol now generally available in Microsoft Foundry (November 17, 2026)
- AWS What's New: OpenAI GPT-6.2 Astra and GPT-6.2 Sol are now generally available on Amazon Bedrock (November 17, 2026)
- RouterPlus model catalog (November 18, 2026)
