GPT-56T 861 —
MUSE-SPK 835 -0.7%
GPT-56SC 828 -5.2%
QWEN-38X 824 —
CL-OP55X 822 —
GROK-46H 822 -5%
GPT-6A 820 —
GLM-5 784 -8.4%
CL-FAB5H 743 -5.6%
KIMI-K3X 742 -8.4%
CL-OP5H 720 -5.8%
CL-OP5X 709 -18%
CL-OP46H 698 -5.9%
CL-OP47H 690 -5.9%
GEM-38FH 677 +0.1%
GEM-37FH 657 -24%
GPT-56S 622 —
CL-OP47 582 -0.7%
GPT-55H 582 —
INKL 531 —
GEM-31P 513 —
GEM-3P 499 —
CL-OP46 496 -0.2%
CL-OP48 490 —
GPT-56T 861 —
MUSE-SPK 835 -0.7%
GPT-56SC 828 -5.2%
QWEN-38X 824 —
CL-OP55X 822 —
GROK-46H 822 -5%
GPT-6A 820 —
GLM-5 784 -8.4%
CL-FAB5H 743 -5.6%
KIMI-K3X 742 -8.4%
CL-OP5H 720 -5.8%
CL-OP5X 709 -18%
CL-OP46H 698 -5.9%
CL-OP47H 690 -5.9%
GEM-38FH 677 +0.1%
GEM-37FH 657 -24%
GPT-56S 622 —
CL-OP47 582 -0.7%
GPT-55H 582 —
INKL 531 —
GEM-31P 513 —
GEM-3P 499 —
CL-OP46 496 -0.2%
CL-OP48 490 —
← Back to feed

OpenAI Configures GPT-6 Astra for Legal Research, Scoring 54% vs 38.7% for Base Model

OpenAI launched Astra for Law on September 17, a vertical configuration of GPT-6 Astra built for legal research and drafting. It is not a new model. It is GPT-6 Astra with a legal search index, custom instructions for legal analysis, and access controls designed for client confidentiality.

What It Includes

Astra for Law pairs GPT-6 Astra with a legal search index covering more than 230 million URLs: U.S. case law, statutes, regulations, court rules, and administrative decisions, updated daily. The case-law collection comes from a partnership with the Free Law Project, the nonprofit behind CourtListener, covering more than 99.9% of published U.S. precedential case law.

Law firms get 26 ecosystem plugins connecting ChatGPT to specialist legal tools including Relativity and Clio. API access is coming soon for customers including Harvey and Legora, who will be able to build on the configuration inside their own products.

The product appears in the ChatGPT model picker as GPT-6 Astra Law and in the API as gpt-6-astra-law. Initial availability is limited to selected law firms through Trusted Access.

Benchmark Results

OpenAI tested the configuration on 200 questions from the private validation set of Vals AI’s Legal Research Bench, which measures both source retrieval and answer quality under evaluation criteria developed for professional legal work.

At highest reasoning effort, Astra for Law passed the overall correctness check on 54.0% of questions. GPT-6 Astra using web search alone passed 38.7% — a 40% relative improvement from the legal search index and tailored instructions alone.

On case-law-focused questions, Astra for Law found 24% more reference cases than the base model. On the audited set of target passages, it retrieved up to 54% more relevant passages from the correct court opinions.

OpenAI also compared against Claude Fable 5.1. In the litigation example used in their testing, Fable 5.1 returned a holding that had been reversed on appeal. In the transactional example, Fable 5.1 reported finding no relevant case. Astra for Law returned matching precedents in both.

The Market

Vertical AI for legal work is contested. Harvey, backed by major law firm partnerships and OpenAI funding, is simultaneously a named API partner and a competitor in the broader legal AI layer. Legora, a European legal tech firm, is also named as an upcoming API customer.

Thomson Reuters joined as a launch partner; its HighQ matter-context integration is among the 26 active plugins, with a CoCounsel Legal connector previewed as a forthcoming addition. Astra for Law is positioned as infrastructure that complements existing legal databases and workflow tools rather than replacing them.