# GPT-6 Astra Law Pricing: $10 Input, $50 Output and API Access Limits

OpenAI lists promotional token rates for its legal configuration, but a pricing entry does not mean the API is generally available.

By Paras Madan | 2026-11-01


OpenAI's API pricing page now lists `gpt-6-astra-law` at $10 per million input tokens, $50 per million output tokens and $1 per million cached input tokens. The listing gives legal technology teams a budgeting reference, but OpenAI still describes API availability as coming soon.

[TestingCatalog highlighted the pricing entry on October 31](https://x.com/TestingCatalog/status/2116443618659455284). The rates are directly verifiable in OpenAI's documentation; the exact time they were added is not established by that observation. Astra for Law itself was introduced on September 17, so this is a pricing development rather than a new model launch.

## GPT-6 Astra Law token pricing

The [official pricing page](https://developers.openai.com/api/docs/pricing) places the model under legal configurations and lists these promotional rates:

| Usage | Price per million tokens |
| --- | --- |
| Input | $10 |
| Output | $50 |
| Cached input | $1 |

OpenAI says the promotional pricing remains available at least through December 31, 2026. That is not a confirmed price increase on January 1, and the entry does not establish an earlier price from which to calculate a discount.

For illustration, a workload with 100,000 uncached input tokens and 10,000 output tokens would cost $1.50 for those tokens at the listed rates. This is arithmetic, not a measured legal research session or an all-in quote. It excludes any other billable usage and should not be treated as the cost of a completed matter.

## Is the Astra Law API available?

**A listed price is not proof of general API access.** OpenAI's [current access guidance](https://help.openai.com/en/articles/20001528-astra-for-law) says Astra for Law is initially for selected U.S. law firms, with access for eligible lawyers and people working under their supervision. Firms can request early access through their account team or OpenAI Sales.

The guidance still describes the API as coming soon. The original announcement identifies `gpt-6-astra-law` as its API name, but an identifier alone does not establish that an arbitrary developer account can call it. Teams planning an integration should confirm access before building a launch schedule around the pricing table.

## What makes Astra for Law different?

OpenAI's [September announcement](https://openai.com/index/astra-for-law/) describes GPT-6 Astra combined with legal analysis instructions and a dedicated search index. The index covers U.S. case law, statutes, regulations, court rules and administrative decisions, with new sources added daily.

That setup matters because the work includes finding relevant authority and connecting it to the facts, rather than only producing fluent legal prose. It does not remove the need to inspect the sources and review the answer. A U.S.-focused index also should not be read as evidence of equivalent coverage in every jurisdiction.

## What teams can budget now

The useful next step is to separate the model bill from the cost of producing a reviewed result. A longer research answer can consume more output tokens while still leaving a lawyer with substantial verification work. Conversely, a higher token bill could be worthwhile if the retrieved authorities are relevant and easy to check.

The pricing entry makes that evaluation easier to plan. The harder question remains how much review a team needs before it can use the output, and that needs evidence from its own authorized workflow rather than a token-rate comparison alone.

I’m Paras. I work at the intersection of AI, engineering and distribution.

I help technical teams explain complex AI products and launch updates clearly, so developers and decision-makers understand what is available and why it matters.

[Book a 30-minute call](https://cal.com/paras-madan/30min)