Standard Compute
← All models
GPT-6 Astra · OpenAI

What does GPT-6 Astra cost to run in a coding agent?

GPT-6 Astra costs about $0.41 per average coding-agent request on our measured traffic, at list prices of $10 per million input and $50 per million output tokens. It scores 58.2% on Terminal-Bench 4.0 in Codex, tied with GPT-6.1 Sol.

List prices: OpenAI API pricing, checked 2026-10-13. Request costs: Standard Compute gateway data, 79,301 coding-agent requests from 75 users, Sep 13 to Oct 12, 2026. Benchmark: Terminal-Bench 4.0 leaderboard. How we compute this

Run GPT-6 Astra in your agent from $19/month

See plans

What GPT-6 Astra costs per coding-agent request

Cost at list priceGPT-6 Astra
Prompt under 100K tokens50% of requests, 87% of input cached$0.15
Prompt of 100K to 272K39% of requests, 88% of input cached$0.39
Prompt over 272K tokens11% of requests, 86% of input cached$1.71
Average request$0.41
100 requests$41.25
Median session16 requests; a session ends after 10 idle minutes$4.49
A month of regular usemedian of 20 users active on 5+ days$583

Each logged request priced at the vendor's list rates for its own token counts, including cache reads and any higher rate for long prompts. Average output: 698 tokens per request.

What to watch for

Astra costs 6.1x GPT-6.1 Sol per request for the same Terminal-Bench 4.0 score. Requests over 272K input tokens (11% of ours) bill at $20 and $75 per million and make up 44% of its cost.

Where GPT-6 Astra sits on the Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index bar chart with GPT-6 Astra (max) highlighted, from artificialanalysis.ai, October 13, 2026
Artificial Analysis Intelligence Index v4.3.2, GPT-6 Astra (max) highlighted, downloaded October 13, 2026. Source: Artificial Analysis (artificialanalysis.ai)

About GPT-6 Astra

GPT-6 Astra is the top tier of OpenAI's GPT-6 family, released September 3, 2026 for the most demanding work. It scores 53 on the Artificial Analysis Intelligence Index (max effort), 58.2% on the Terminal-Bench 4.0 leaderboard in Codex, and ranks #2 on LMArena's WebDev board behind Claude Opus 5.5. It has a 1,050,000-token context window (922K input) and 128K output. At $10 per million input tokens it costs five times GPT-6.1 Sol, which OpenAI describes as delivering near-Astra performance, so Astra is mainly worth it when the extra margin on a hard task matters.

Sources: OpenAI model page · Artificial Analysis · Terminal-Bench 4.0 · SWE-Bench Pro V2

Agents that run GPT-6 Astra well

The hardest coding and research tasks in OpenAI-based stacks, where cost is secondary.

GPT-6 Astra FAQ

How much does GPT-6 Astra cost per coding-agent request?
$0.41 on average at list price, $41.25 per 100 requests, and $4.49 for a median 16-request session. That is 79,301 real coding-agent requests (Sep 13 to Oct 12, 2026) priced at OpenAI's list rates; the median request sent 99,534 input tokens, 87% of input was read from cache, and requests returned 698 output tokens on average.
Does GPT-6 Astra charge more for long prompts?
Yes. Prompts over 272K tokens are billed at $20 per million input and $75 per million output tokens, on every token in the request including cache reads. 11% of the coding-agent requests we measured were that large.
Is GPT-6 Astra good for AI agents?
The hardest coding and research tasks in OpenAI-based stacks, where cost is secondary. On our editorial scoring it rates 9.7/10 for agentic ability and 9.7/10 for output quality. Top OpenAI scores: Artificial Analysis 53, Terminal-Bench 4.0 58.2%.
What is GPT-6 Astra's context window?
GPT-6 Astra has a 1.1M-token context window — large enough to hold sizeable codebases or document sets in a single run. OpenAI is the maker.
Can I self-host GPT-6 Astra?
No — GPT-6 Astra is a proprietary, hosted-only model from OpenAI. You access it through an API (e.g. OpenRouter) rather than running the weights yourself.
What are the downsides of GPT-6 Astra?
The main trade-offs: Five times GPT-6.1 Sol's price for a small measured gain; Slow: about 66 output tokens/s at max effort; Prompts over 272K tokens are billed at higher rates. It's strongest for the hardest coding and research tasks in OpenAI-based stacks, where cost is secondary.

GPT-6 Astra compared

The agent matters as much as the model. Which agent gets the most out of GPT-6 Astra? →

Every model on this page is included with Standard Compute

Standard Compute

Current frontier models Claude Opus 5.5 · Claude Sonnet 5.5

Your agents never stop.
Your bill never grows.

Frontier models when it counts. Efficient models when it doesn’t. The most intelligence per dollar. One flat bill.

Plans from $19/mo · cancel anytime · 7-day fair refund

7-day fair refund No credit card needed Billing by Stripe1,800+ agent users