Trending Now
  • Which AI Model Is Best for Legal Work? What 2026 Research Says About Accuracy
  • ClaimAngel Reports $144M Deployed and 30,000 Fundings on Consumer Legal Funding Marketplace
Commercial

Which AI Model Is Best for Legal Work? What 2026 Research Says About Accuracy

September 24, 2026

Which AI Model Is Best for Legal Work? What 2026 Research Says About Accuracy

Law firms, funders and legal departments are being sold AI for contract review, legal research and citation checking, and the models change every few months. So we looked only at independent studies published in 2026 that tested the current generation of models from OpenAI, Anthropic and Google on real legal tasks. The short version: the best models are now genuinely good at reading and extracting from documents you give them, still unreliable at recalling law from memory, and the commercial legal research tools lag behind the best custom systems.

At a Glance

Best overall model for legal document work: Google's Gemini 3.1 Pro. It was at or near the top in every 2026 study that tested it, and it was usually the fastest and cheapest of the leaders. OpenAI's GPT-5.5 found slightly more errors in contract review, and Anthropic's Claude models were the most careful about not flagging problems that weren't there.

Best accuracy recorded on a full legal task: 92%, on a 50-state statutory research test run by Stanford, achieved by a purpose-built research tool. The lesson is that how the AI is set up matters as much as which model sits underneath it.

Range of accuracy: from under 7% (asking a model to recall exact case citations from memory) to 99–100% (catching a citation to the wrong case when the model can read the source). Most real-world document tasks landed between 60% and 85%.

Westlaw and Lexis AI: 58% and 64% accuracy on a Stanford statutory survey test, below a custom-built tool at 83–92%.

Biggest single improvement: giving the model the actual documents instead of asking it from memory cut fabricated citations from roughly 15–40% to about 4–15%, and to under 0.2% with a well-built retrieval system.

Key Takeaways

The model to use. For contract review and extraction, start with Gemini 3.1 Pro. It matched the top performer on catching contract errors (74% vs. 75%) at about one-seventh of the cost and in 90 seconds instead of nine minutes. If catching every possible issue matters more than time or cost, GPT-5.5 with reasoning turned on found the most. For checking citations in a brief, the best 2026 results came from GPT-5 running as an agent and from Claude Code with Claude Opus, which was the most precise.

How to Use It

  • Give it the documents. Never ask a model to supply case law or citations from memory.
  • Turn on the model's "reasoning" or "thinking" mode for review work. It improved error-catching by 9 to 11 points in contract review.
  • Use it as a first pass and a second reviewer, not the final reviewer.
  • For research, use a tool that pulls from a full, current database of the law, because weak retrieval, not the model, causes many of the errors.

What to Expect

  • Contract extraction (pulling out dates, parties, termination and liability terms): about 80–84% accuracy for the best models.
  • Final contract proofreading (defined terms, cross-references, inconsistent language): the best models catch about three-quarters of errors. On a 60-page agreement, expect it to miss some.
  • Citation checking: nearly all citations to the wrong case get caught, but wrong pinpoint pages slip through 20% to 60% of the time.
  • Research answers grounded in the right documents: roughly 6–11% of answers still contain an unsupported statement.

What to Look Out For

  • Citations from memory. When asked to recall exact citations without sources, even the best model scored under 7 out of 100, and 20 of 21 models gave confident, wrong answers more than 94% of the time.
  • Right case, wrong page. Models tend to approve a citation because the case is on the right topic, even when the cited page doesn't support the point.
  • Questions with a false premise. If your question assumes something that isn't true, models often go along with it.
  • Legal research tools' marketing. Westlaw AI and Lexis+ AI trailed a custom-built tool by 19 to 25 points on a Stanford test.
  • Studies funded by vendors. Some of the best-looking results come from companies selling legal AI. Check who ran the test.

Best Practices

  • Ground every task in source documents, and require the model to quote the passage it relied on.
  • Check every citation yourself at the pinpoint page before filing. Automated checkers help but don't replace this.
  • Turn on reasoning mode for review tasks and accept that it's slower.
  • Test a model on a few of your own documents before rolling it out. Rankings change by task.
  • Re-test when a new model version arrives; this field moves in months, not years.
  • Keep a human reviewer accountable for anything that goes to a court, a client or a counterparty.

What the Studies Found

Contract proofreading. In August 2026, researchers had experienced lawyers plant errors in contracts (misused defined terms, wrong cross-references, wrong party names, contradictions) and tested ten current models on catching them. GPT-5.5 caught 75% of errors, Gemini 3.1 Pro 74%, Claude Sonnet 4.6 69% and Claude Opus 4.7 62%. GPT-5.5 cost $1.38 per contract and took about nine minutes; Gemini 3.1 Pro cost $0.19 and took about 90 seconds. Turning on reasoning mode added 9 to 11 points. Every model was far cheaper than a lawyer, and none was close to perfect.

Contract extraction. A May 2026 study tested models on pulling 26 standard fields out of contracts. Among the major models, Gemini 3.1 Pro scored highest (82%), with Claude Opus 4.6 (82%) and Claude Sonnet 4.6 (80%) close behind and GPT-5.4 at 78%. A smaller legal-specific model built by the study's authors scored 84% at far lower cost. The authors work for Onit, which makes that model.

Made-up citations and facts. A January 2026 study had expert reviewers check 2,700 legal answers from 12 models. Asked without source documents, the best models (GPT-5.2 and Gemini 3.0 Pro) cited something false about 15–17% of the time, and the worst over 30%. Giving the models the relevant documents cut that to about 4–15%. A more carefully built retrieval system brought it below 0.2% for every model.

Research with sources. A March 2026 study found that when models answer from retrieved legal texts, Gemini 3.1 Pro produced unsupported statements 5.7% of the time versus 11.3% for GPT-5.2, and that the quality of the search step mattered more than the choice of model. Its authors sell the search component that performed best. An August 2026 study of eight research setups found unsupported answers ranging from under 10% for the best to nearly half for the worst, with the worst results on questions built on a false assumption.

Westlaw and Lexis. In a February 2026 Stanford study, researchers tested legal AI tools against a Department of Labor survey of state unemployment insurance laws. Westlaw AI scored 58% and Lexis+ AI 64%, while a custom statutory research tool scored 83%, rising to 92% after the researchers found that some of its "errors" were gaps in the government's own survey.

Citation checking. A June 2026 study found more than 1,000 court filings containing fabricated citations, a number growing every year, and tested AI checkers on catching them. GPT-5, working as an agent that looks up cases, caught 83% of planted errors; Claude Code running Claude Opus 4.8 was the most precise and scored best overall. No model reliably caught wrong pinpoint cites, partly because page numbers often sit behind Westlaw and Lexis paywalls. A separate August 2026 study found models catch 93–100% of citations to the wrong case but miss many citations to the wrong page, and even GPT-5.4 with full reasoning missed 40% of wrong pinpoints in court opinions.

Citations from memory. A May 2026 study built from 1,000 real U.S. judicial opinions asked 21 models to recall exact case citations without any sources. The best, Claude Sonnet 4.5, scored under 7 out of 100.

The Bottom Line

The 2026 research is consistent: today's best models, led by Gemini 3.1 Pro, GPT-5.5 and Claude, are useful and cheap for first-pass contract review and extraction when they work from the documents in front of them. They still invent law when asked from memory and still miss wrong pinpoint citations, so a lawyer has to verify anything that leaves the building.

Sources (All 2026)

  • Bang et al., "ContractScrub: A benchmark for final review of legal contracts" (Aug. 2026), arXiv:2608.20204
  • Lincoln et al., "A Few Good Clauses: Comparing LLMs vs Domain-Trained Small Language Models on Structured Contract Extraction" (May 2026), arXiv:2605.05532
  • Dantart, "Reliability by design: quantifying and eliminating fabrication risk in LLMs" (Jan. 2026), arXiv:2601.15476
  • Butler and Butler, "Legal RAG Bench: an end-to-end benchmark for legal RAG" (Mar. 2026), arXiv:2603.01710
  • Das et al., "How Much Do Legal RAG Systems Still Hallucinate?" (Aug. 2026), arXiv:2608.14210
  • Afane et al., "Benchmarking Legal RAG: The Promise and Limits of AI Statutory Surveys" (Feb. 2026), arXiv:2603.03300
  • Liu, Stammbach and Henderson, "Who Checks the Citations? Benchmarking Legal Hallucination Detection" (June 2026), arXiv:2606.21155
  • Verma, "Is this Citation on Point?" (Aug. 2026), arXiv:2608.12571
  • Chen et al., "LegalCiteBench: Evaluating Citation Reliability in Legal Language Models" (May 2026), arXiv:2605.10186

Commercial

View All
Commercial

Second Circuit Affirms Fee Award That Treated Litigation Funding Costs as Firm Overhead

The Second Circuit has upheld a $4.8 million attorneys' fee award in a sex trafficking case, endorsing a district court's decision to strike time counsel spent communicating with its litigation funder.

In Moore v. Rubin, decided on September 4, a panel of Chief Judge Lohier and Judges Parker and Chin affirmed the award to six plaintiffs who won a $3.85 million jury verdict against former bond trader Howard Rubin under the Trafficking Victims Protection Act. In rejecting the argument that too many timekeepers had been compensated, the panel noted approvingly that the district court had applied a 15% across-the-board reduction and excluded non-compensable tasks, "such as communications with counsel's litigation funder."

The more consequential ruling for funders came below. In February 2025, Judge Brian Cogan of the Eastern District of New York refused to shift roughly $1.84 million in principal and interest owed to a third-party funder, reasoning that how a lawyer finances a practice is irrelevant to the client and the defendant alike. "Whether it is a bank loan, family loan, personal assets, or a litigation funder," he wrote, "it is overhead."

Judge Cogan also declined to follow the English decision in Essar Oilfield Services v. Norscot Rig Management, which allowed recovery of funding costs, observing that neither the statute nor the local rule hints at such recovery.

The funding cost denial was not before the appellate panel, as Rubin appealed only the fee award. The funder, Pravati Investment Fund IV, later sought unsuccessfully to intervene to protect its interest in the fees after the plaintiffs' firm dissolved.

Commercial

Legalist Asks Manhattan Federal Court to Confirm $108,718 Award Against Funded Claimant

Litigation funder Legalist has asked a federal court in Manhattan to confirm an arbitration award against a claimant it financed, in a rare public dispute between a funder and the plaintiff whose case it paid for.

As reported by Bloomberg Law, Legalist filed its petition on September 8 in the Southern District of New York, seeking to confirm a partial final award of $108,718.45 against Mario Rinaldi. The award, issued on May 18 by JAMS arbitrator the Hon. Elaine Rushing (Ret.), consists of $105,218.45 in attorneys' fees and $3,500 in arbitration costs.

The funding agreement dates to June 2018 and financed Rinaldi's suit against two French champagne producers, which he brought after working to build their brand in New York. A jury returned a $1.5 million verdict in his favor in March 2022, and final judgment with prejudgment interest was entered that December at $2,318,506.85. That judgment has not been collected. Rinaldi told Bloomberg Law he is still pursuing the money with his own resources, having retained French counsel to enforce it abroad.

Legalist alleges in its filing that Rinaldi breached the funding agreement by refusing to cooperate in collection efforts, including by declining to permit communication with his French counsel. According to the memorandum, Rinaldi did not appear in the arbitration at any stage, and has not moved to vacate or modify the award.

The arbitration was held open for Legalist to pursue further relief, indicating the $108,718.45 may not represent the full extent of its claim. No response from Rinaldi appears in the most recent public docket entries.

Commercial Consumer

ClaimAngel Reports $144M Deployed and 30,000 Fundings on Consumer Legal Funding Marketplace

South Florida consumer legal funding marketplace ClaimAngel says it has deployed more than $144 million across over 30,000 individual fundings since launching in April 2023, positioning standardised pricing as its answer to the cost criticisms that dog the consumer funding sector.

As reported by Refresh Miami, the platform runs a marketplace in which 27 funding providers compete for cases, with funders reserving a case in an average of 11 seconds. Advances carry 27.8% simple interest with no compounding and a 2x cap on total repayment, and remain non-recourse — plaintiffs owe nothing if the case is lost.

The company reports serving more than 14,500 plaintiffs and over 750 law firms, with 46 employees. A Case Equity product lets plaintiffs draw against expected case value for living expenses while litigation is pending.

ClaimAngel was co-founded by Jeremy Alters, a trial lawyer of more than two decades who was disbarred by the Florida Supreme Court in 2018 for misusing client funds, and his son Logan Alters. "I did things wrong. They were my fault. I take full responsibility for it," Jeremy Alters said, describing the company as "born out of an ethics issue." He applied for readmission to the Florida Bar in 2025.

Planned expansions include attorney funding, a secondary marketplace for buying and selling existing positions, and AngelScore, a data-driven underwriting system.