Home › Best AI by category › Drafting › Report
As of Oct 26, 2026, aiforlawyers.app counts 1033 lawyer voices on drafting; 925 voices name one of 14 tools, 8 of which have 10+ voices and are rated. Highest floor of the 90% band: Claude (net +57, band +47 to +65, 141 voices, Medium confidence).
Download the PDF (cutoff Oct 26, 2026)Data (JSON)Ranking page
Cite this
aiforlawyers.app (AI Vortex), Drafting report, cutoff Oct 26, 2026, https://aiforlawyers.app/report/drafting, licensed CC BY 4.0.
For one figure: “As of Oct 26, 2026, aiforlawyers.app counts 1033 lawyer voices on drafting; 925 voices name one of 14 tools, 8 of which have 10+ voices and are rated. Highest floor of the 90% band: Claude (net +57, band +47 to +65, 141 voices, Medium confidence).”
14 tools named in 1033 voices on drafting. Rated tools first, ordered by the floor of the 90% band; then the tools with few voices, with their count and no band. Net = like minus dislike per 100 voices.
| # | Tool | Status | Voices | Net | 90% band | Confidence | Top complaint (voices) |
|---|---|---|---|---|---|---|---|
| 1 | Claude | Rated | 141 | +57 | +47 to +65 | Medium | Accuracy (17) |
| 2 | Gemini | Rated | 27 | +48 | +20 to +68 | Low | Accuracy (5) |
| 3 |  ChatGPT | Rated | 573 | +22 | +16 to +28 | High | Accuracy (120) |
| 4 |  | Rated | 22 | +32 | +5 to +53 | Low | Accuracy (5) |
| 5 |  Harvey | Rated | 59 | +20 | +3 to +36 | Low | Accuracy (13) |
| 6 |  Copilot | Rated | 30 | +17 | −8 to +40 | Low | Accuracy (7) |
| 7 | CoCounsel | Rated | 35 | +3 | −22 to +27 | Low | Accuracy (16) |
| 8 | Lexis+ AI | Rated | 18 | −11 | −42 to +22 | Low | Accuracy (8) |
| – |  Legora | Few voices | 9 | – | no band: 9 of 10 voices needed | not rated | Accuracy (3) |
| – | Perplexity | Few voices | 4 | – | no band: 4 of 10 voices needed | not rated | Accuracy (1) |
| – | Westlaw AI | Few voices | 3 | – | no band: 3 of 10 voices needed | not rated | Trust, liability (1) |
| – |  Clio | Few voices | 2 | – | no band: 2 of 10 voices needed | not rated | none on record |
| – |  Luminance | Few voices | 1 | – | no band: 1 of 10 voices needed | not rated | Price (1) |
| – |  | Few voices | 1 | – | no band: 1 of 10 voices needed | not rated | none on record |
The most named complaints per tool, with a verbatim quote each. Counts are voices, not opinions of the firm or of us.
34 voices of 141 criticize it or are mixed about it with a named complaint: accuracy (17), hallucination (8), trust, liability (3), integration (3), price (2), privacy (1).
“This seems like a particularly BAD use case for generic Claude right now.”
“Working on a legal motion, starts making shit up, pure fiction, not in the knowledge base.”
8 voices of 27 criticize it or are mixed about it with a named complaint: accuracy (5), hallucination (3).
“I had not used #gemini in awhile but had good results with it in the past for complex tasks...”
“I decided I’m going to read that case to understand the context of the case. It was completely fake.”
ChatGPTRated · 573 voices244 voices of 573 criticize it or are mixed about it with a named complaint: accuracy (120), hallucination (93), privacy (15), trust, liability (13), price (2), integration (1).
“More annoyingly, I’ll write really thought-out memos and he “revises” them by plugging them into AI and generating a more apathetic, em dash filled version that at least IMO is not better.”
“RFK Jr’s HHS getting caught submitting legal filings where 22 out of 29 footnotes were fabricated by AI!”
Rated · 22 voices9 voices of 22 criticize it or are mixed about it with a named complaint: accuracy (5), hallucination (2), trust, liability (1), price (1).
“It is unbelievably janky and the drafting quality/consistency is frankly poor.”
“I have used Spellbook, which I think is probably solid solution for transactional attorneys, but for litigation, it’s only slightly better than ChatGPT (hallucinations are common) and for a much higher price point, so I paused my subscription.”
HarveyRated · 59 voices29 voices of 59 criticize it or are mixed about it with a named complaint: accuracy (13), hallucination (6), price (4), trust, liability (3), integration (3).
“Hard to see how that is really worth the cost savings of just having us do it right form the start.”
“It hallucinated a board action with citation to the wrong DGCL section.”
CopilotRated · 30 voices12 voices of 30 criticize it or are mixed about it with a named complaint: accuracy (7), hallucination (3), privacy (1), integration (1).
“I was working with someone who is fairly neutral on AI and was just incensed that Copilot couldn’t make a template for legal numbering.”
“The output *looks* professional and plausible. However, I found errors in legal explanations, citations that were nonexistent, and different responses to the same prompt.”
20 voices of 35 criticize it or are mixed about it with a named complaint: accuracy (16), integration (2), privacy (1), price (1).
“My boss used it in a conclusion for a brief, and it called the same party 2 different things and did not accurately summarize the argument.”
“The goal is for CoCounsel to replace it because we can connect CoCounsel to other things I'm building in Claude but the current set up of CoCounsel is choking on file uploads and has too many restricted file types to be particularly useful in the document heavy cases we work.”
11 voices of 18 criticize it or are mixed about it with a named complaint: accuracy (8), price (2), integration (1).
“I can’t stand Lexis AI for writing. I have not tried Claude.”
“I think ai from the existing big players is a mistake. Costs more than it should, they don't specialize, they generalize with items you don't need.”
LegoraFew voices · 9 of 104 voices of 9 criticize it or are mixed about it with a named complaint: accuracy (3), integration (1).
“I love Legora for those purposes, first drafts, brainstorming. But these AI solutions are not reliable”
“I can't stand the card-based Word Add In interface.”
1 voice of 4 criticize it or are mixed about it with a named complaint: accuracy (1).
“I got a near final doc from one the other day with a bunch of perplexity bullshit in it.”
1 voice of 3 criticize it or are mixed about it with a named complaint: trust, liability (1).
“I don’t recommend the AI stuff that’s out there, at least not yet.”
ClioFew voices · 2 of 10No named complaint on record among its 2 voices.
LuminanceFew voices · 1 of 101 voice of 1 criticize it or are mixed about it with a named complaint: price (1).
“It’s also an absolute beast to set up, but once you get there it is incredibly powerful and the UI is slick.”
Few voices · 1 of 10No named complaint on record among its 1 voice.
Public tests that include drafting work, with their own publisher, date and caveat. They measure question sets, not lawyers’ experience, and are never added to the ranking above. Others are on the arena.
drafting-specific independent test · Legal Benchmarks (Anna Guo and co-authors)
published 2025-09 (preliminary findings; tools run July to August 2025) · read 2026-10-23
30 contract-drafting tasks (basic clauses, template adaptation, bespoke clauses; junior to mid level) written by practising lawyers, run on 13 AI tools plus in-house commercial lawyers (about 10 years' experience) as a baseline, 450 task outputs in all. A draft passed only if it met every key element from the contributing lawyer's answer key plus instruction compliance, factual accuracy and legal adequacy. Pass rate averaged 57.3% across the AI tools; the human lawyers reached 56.7%. Top individual tool: Gemini 2.5 Pro at 73.3%, against 70% for the best individual lawyer. The authors report that legal-specific tools did not clearly beat general assistants on draft quality, and that legal tools raised explicit risk warnings more often (83% of high-risk outputs versus 55% for general tools).
Caveat. Run by an independent project, not by a vendor, but the site also sells benchmarking and some tool descriptions carry affiliate-tagged links; 30 tasks and one prompt per tool with default features, so a snapshot of 2025 versions that is over a year old; an LLM jury graded reliability with expert review of borderline cases, and usefulness and workflow scores are expert opinion; one tool is anonymized; the authors say their 2025 figures are not comparable with the current leaderboard pass rates; read from the live page 2026-10-23.
Names: Gemini 2.5 Pro, GPT-5, GC AI, Brackets, August, SimpleDocs.
Related on AI Vortex
Use these numbers: the counts, rates and bands on this report are free to reuse under CC BY 4.0. Credit “aiforlawyers.app (AI Vortex)” with the date and a link to the page. Quotes belong to the people who wrote them: quote them with the link to the original post. Machine-readable: index.json. How to cite.
