How it changed, quarter by quarter.
Sentiment and volume per tool and per model, from 2023 Q1 to now, each cell with its n and band, and every marked change next to the events around it. As of Oct 26, 2026.
Marked changes
A change is marked only when the 90% bands of two consecutive quarters do not overlap (net, or share of all voices on named tools). 26 marked.
- CoCounsel · 2026 Q4* · share of voice 6% → 14% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Claude · 2026 Q4* · share of voice 36% → 27% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Sep 28, 2026: Anthropic introduces Claude Sonnet 5.5 sourceOct 7, 2026: Anthropic introduces Claude Haiku 5.5 sourceSep 22, 2026: Anthropic introduces Claude Opus 5.5 source
- Spellbook · 2026 Q2 · share of voice 3% → 1% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Lexis+ AI · 2026 Q2 · share of voice 3% → 6% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Claude · 2026 Q2 · share of voice 17% → 36% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jun 30, 2026: Anthropic releases Claude Sonnet 5 source
- ChatGPT · 2026 Q2 · share of voice 35% → 17% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Legora · 2026 Q1 · share of voice 4% → 1% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Mar 10, 2026: Legora announces $550M Series D at $5.55B valuation source
- Harvey · 2026 Q1 · share of voice 13% → 18% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- CoCounsel · 2026 Q1 · net +32 → −21 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Clio · 2026 Q1 · share of voice 4% → 1% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Nov 10, 2025: Clio closes US$1B acquisition of vLex and announces Series G source
- Claude · 2026 Q1 · share of voice 5% → 17% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Nov 24, 2025: Anthropic releases Claude Opus 4.5 source
- ChatGPT · 2026 Q1 · share of voice 45% → 35% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- Legora · 2025 Q4 · share of voice 1% → 4% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 23, 2025: N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations source
- Harvey · 2025 Q4 · net −46 → −15 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 23, 2025: N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations source
- Clio · 2025 Q4 · share of voice 1% → 4% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Nov 10, 2025: Clio closes US$1B acquisition of vLex and announces Series G sourceJul 23, 2025: N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations source
- Harvey · 2025 Q3 · share of voice 7% → 15% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 23, 2025: N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations sourceJun 6, 2025: England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun source
- Clio · 2025 Q3 · net −100 → +40 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 23, 2025: N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations sourceJun 6, 2025: England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun source
- Lexis+ AI · 2025 Q2 · share of voice 13% → 6% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jun 6, 2025: England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun source
- Copilot · 2025 Q2 · net −64 → +25 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jun 6, 2025: England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun source
- Copilot · 2025 Q2 · share of voice 6% → 2% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jun 6, 2025: England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun source
- Copilot · 2025 Q1 · net +33 → −64 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- ChatGPT · 2024 Q4 · net −8 → +22 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 29, 2024: ABA issues Formal Opinion 512 on generative AI tools source
- Lexis+ AI · 2024 Q3 · share of voice 4% → 11% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jul 29, 2024: ABA issues Formal Opinion 512 on generative AI tools sourceMay 23, 2024: Stanford RegLab/HAI preprint on hallucination rates of legal AI research tools source
- ChatGPT · 2024 Q1 · share of voice 83% → 63% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):Jan 23, 2024: Florida Bar news reports Board of Governors adopts generative AI ethics guidelines sourceNov 16, 2023: State Bar of California publishes generative AI practical guidance source
- ChatGPT · 2023 Q3 · net −40 → −9 vs the previous quarter (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
- ChatGPT · 2023 Q2 · share of voice 98% → 90% (90% bands do not overlap)Events in the window (the quarter and the one before; context, not proof of cause):No event in our verified list for this window.
Sentiment by quarter
Net (like − dislike) per tool and quarter, with its 90% band, from 5 voices; below that only the count.
| Tool | 2023 Q1 | 2023 Q2 | 2023 Q3 | 2023 Q4 | 2024 Q1 | 2024 Q2 | 2024 Q3 | 2024 Q4 | 2025 Q1 | 2025 Q2 | 2025 Q3 | 2025 Q4 | 2026 Q1 | 2026 Q2 | 2026 Q3 | 2026 Q4* |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ChatGPT | −35−45 to −24 n=160 | −40−48 to −31 n=243 | −9−24 to +7 n=91 | −4−21 to +14 n=78 | −9−24 to +6 n=92 | −4−19 to +10 n=93 | −8−22 to +6 n=103 | +22+7 to +36 n=106 | +2−11 to +15 n=122 | −6−15 to +3 n=273 | −11−19 to −3 n=309 | −6−16 to +4 n=221 | −19−28 to −10 n=262 | −14−26 to −1 n=136 | −16−28 to −3 n=134 | −32−51 to −9 n=44 |
| Claude | n=0 | n=0 | n=0 | n=1 | n=2 | +64+16 to +87 n=11 | +62+9 to +85 n=8 | +100+36 to +99 n=5 | +56+21 to +76 n=16 | +28−1 to +52 n=29 | +71+51 to +82 n=31 | +56+30 to +72 n=25 | +28+15 to +40 n=124 | +33+25 to +41 n=284 | +34+27 to +42 n=285 | +49+31 to +63 n=51 |
| Harvey | n=1 | +50−9 to +80 n=6 | n=1 | n=2 | n=3 | −50−74 to −11 n=14 | n=1 | −18−57 to +25 n=11 | n=4 | −36−56 to −10 n=36 | −46−58 to −32 n=93 | −15−29 to +2 n=62 | −12−24 to 0 n=136 | −25−36 to −14 n=168 | −12−22 to 0 n=165 | −20−38 to 0 n=46 |
| CoCounsel | n=2 | +59+28 to +76 n=17 | n=3 | +20−16 to +51 n=5 | −17−54 to +25 n=12 | +15−15 to +42 n=13 | +9−19 to +35 n=23 | −23−56 to +19 n=13 | +29−4 to +54 n=14 | +22+1 to +41 n=45 | −11−32 to +11 n=45 | +32+7 to +53 n=34 | −21−37 to −3 n=67 | −27−45 to −6 n=44 | +6−14 to +25 n=47 | +4−17 to +25 n=27 |
| Lexis+ AI | n=0 | n=0 | n=1 | n=1 | −11−45 to +29 n=9 | −29−61 to +17 n=7 | +10−23 to +40 n=21 | −42−60 to −19 n=38 | 0−25 to +25 n=31 | −9−32 to +15 n=32 | −27−54 to +5 n=22 | +19−8 to +42 n=26 | −5−35 to +26 n=22 | −6−26 to +14 n=47 | −49−65 to −25 n=35 | n=2 |
| Westlaw AI | n=0 | n=1 | n=1 | n=1 | −50−72 to −11 n=12 | +8−29 to +41 n=13 | +9−20 to +36 n=22 | −11−40 to +20 n=18 | +11−28 to +45 n=9 | +10−13 to +31 n=31 | +3−20 to +26 n=34 | +9−15 to +31 n=35 | +13−10 to +34 n=39 | +12−17 to +39 n=26 | +5−27 to +34 n=21 | 0−52 to +53 n=5 |
| Copilot | n=0 | n=1 | n=0 | n=0 | n=2 | n=2 | +20−34 to +61 n=5 | +33−14 to +67 n=9 | −64−81 to −29 n=14 | +25−23 to +60 n=8 | −5−35 to +24 n=19 | +19−11 to +46 n=21 | +13−13 to +37 n=30 | −5−36 to +28 n=20 | +26−5 to +51 n=19 | 0−53 to +52 n=5 |
| Gemini | n=0 | n=1 | n=0 | n=4 | −50−80 to +11 n=6 | n=3 | n=3 | n=1 | −22−57 to +21 n=9 | +48+17 to +70 n=21 | +61+29 to +80 n=18 | +50+18 to +70 n=18 | +29−8 to +59 n=17 | +36+5 to +60 n=22 | −10−46 to +31 n=10 | n=1 |
| Spellbook | n=1 | n=0 | n=2 | n=1 | +50−10 to +80 n=6 | +40−23 to +77 n=5 | n=3 | +25−10 to +52 n=12 | +67+15 to +85 n=6 | −17−54 to +28 n=6 | +46+3 to +74 n=13 | n=4 | +52+22 to +72 n=21 | +60+6 to +81 n=5 | +31−11 to +64 n=13 | n=1 |
| Legora | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=2 | +29−16 to +61 n=7 | −5−34 to +26 n=19 | −50−74 to −8 n=10 | +11−21 to +39 n=18 | −7−35 to +21 n=28 | −17−62 to +37 n=6 |
| Clio | n=0 | n=0 | n=0 | n=0 | n=0 | n=2 | n=0 | n=0 | n=0 | −100−99 to −42 n=6 | +40−22 to +77 n=5 | −42−68 to −5 n=19 | +17−37 to +63 n=6 | −6−32 to +22 n=17 | 0−28 to +28 n=21 | n=2 |
| Perplexity | n=0 | n=0 | n=0 | n=0 | n=1 | n=1 | n=1 | n=4 | n=3 | n=2 | +100+34 to +99 n=5 | n=2 | +570 to +83 n=7 | n=0 | +60+17 to +81 n=10 | n=0 |
| vLex Vincent | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=1 | n=2 | n=3 | +80+19 to +92 n=5 | n=2 | +43−17 to +80 n=7 | 0−53 to +51 n=5 | +40−25 to +75 n=5 | n=0 | n=1 |
| Paxton | n=0 | n=0 | n=0 | n=1 | n=1 | n=1 | n=2 | n=1 | n=3 | n=0 | n=0 | n=1 | n=0 | n=1 | n=1 | n=0 |
| Luminance | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | −12−55 to +37 n=8 | n=0 | n=0 | n=0 | n=1 | n=0 | n=0 | n=0 | n=0 |
Volume by quarter, as share of voice
Share of all voices on named tools in that quarter, with a 90% band. Shares travel better than counts: how much we could collect differs by period and source.
| Tool | 2023 Q1 | 2023 Q2 | 2023 Q3 | 2023 Q4 | 2024 Q1 | 2024 Q2 | 2024 Q3 | 2024 Q4 | 2025 Q1 | 2025 Q2 | 2025 Q3 | 2025 Q4 | 2026 Q1 | 2026 Q2 | 2026 Q3 | 2026 Q4* |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ChatGPT | 98%95–99% n=160 | 90%87–93% n=243 | 92%87–96% n=91 | 83%76–89% n=78 | 63%56–69% n=92 | 56%50–62% n=93 | 53%48–59% n=103 | 46%41–52% n=106 | 52%47–57% n=122 | 55%51–59% n=273 | 51%48–55% n=309 | 45%41–48% n=221 | 35%32–38% n=262 | 17%15–19% n=136 | 17%15–19% n=134 | 23%18–29% n=44 |
| Claude | n=0 | n=0 | n=0 | 1%0–4% n=1 | 1%0–4% n=2 | 7%4–10% n=11 | 4%2–7% n=8 | 2%1–4% n=5 | 7%4–10% n=16 | 6%4–8% n=29 | 5%4–7% n=31 | 5%4–7% n=25 | 17%14–19% n=124 | 36%33–39% n=284 | 36%33–39% n=285 | 27%22–32% n=51 |
| Harvey | 1%0–2% n=1 | 2%1–4% n=6 | 1%0–4% n=1 | 2%1–6% n=2 | 2%1–5% n=3 | 8%5–13% n=14 | 1%0–2% n=1 | 5%3–8% n=11 | 2%1–4% n=4 | 7%6–9% n=36 | 15%13–18% n=93 | 13%10–15% n=62 | 18%16–21% n=136 | 21%19–24% n=168 | 21%19–23% n=165 | 24%19–29% n=46 |
| CoCounsel | 1%0–3% n=2 | 6%4–9% n=17 | 3%1–7% n=3 | 5%2–10% n=5 | 8%5–13% n=12 | 8%5–12% n=13 | 12%9–16% n=23 | 6%4–9% n=13 | 6%4–9% n=14 | 9%7–11% n=45 | 7%6–9% n=45 | 7%5–9% n=34 | 9%7–11% n=67 | 6%4–7% n=44 | 6%5–7% n=47 | 14%10–19% n=27 |
| Lexis+ AI | n=0 | n=0 | 1%0–4% n=1 | 1%0–4% n=1 | 6%3–10% n=9 | 4%2–7% n=7 | 11%8–15% n=21 | 17%13–21% n=38 | 13%10–17% n=31 | 6%5–8% n=32 | 4%3–5% n=22 | 5%4–7% n=26 | 3%2–4% n=22 | 6%5–7% n=47 | 4%3–6% n=35 | 1%0–3% n=2 |
| Westlaw AI | n=0 | 0%0–1% n=1 | 1%0–4% n=1 | 1%0–4% n=1 | 8%5–12% n=12 | 8%5–12% n=13 | 11%8–16% n=22 | 8%5–11% n=18 | 4%2–6% n=9 | 6%5–8% n=31 | 6%4–7% n=34 | 7%5–9% n=35 | 5%4–7% n=39 | 3%2–4% n=26 | 3%2–4% n=21 | 3%1–5% n=5 |
| Copilot | n=0 | 0%0–1% n=1 | n=0 | n=0 | 1%0–4% n=2 | 1%0–3% n=2 | 3%1–5% n=5 | 4%2–6% n=9 | 6%4–9% n=14 | 2%1–3% n=8 | 3%2–4% n=19 | 4%3–6% n=21 | 4%3–5% n=30 | 3%2–4% n=20 | 2%2–3% n=19 | 3%1–5% n=5 |
| Gemini | n=0 | 0%0–1% n=1 | n=0 | 4%2–9% n=4 | 4%2–8% n=6 | 2%1–4% n=3 | 2%1–4% n=3 | 0%0–2% n=1 | 4%2–6% n=9 | 4%3–6% n=21 | 3%2–4% n=18 | 4%2–5% n=18 | 2%2–3% n=17 | 3%2–4% n=22 | 1%1–2% n=10 | 1%0–2% n=1 |
| Spellbook | 1%0–2% n=1 | n=0 | 2%1–6% n=2 | 1%0–4% n=1 | 4%2–8% n=6 | 3%1–6% n=5 | 2%1–4% n=3 | 5%3–8% n=12 | 3%1–5% n=6 | 1%1–2% n=6 | 2%1–3% n=13 | 1%0–2% n=4 | 3%2–4% n=21 | 1%0–1% n=5 | 2%1–3% n=13 | 1%0–2% n=1 |
| Legora | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | 0%0–1% n=2 | 1%1–2% n=7 | 4%3–5% n=19 | 1%1–2% n=10 | 2%2–3% n=18 | 4%3–5% n=28 | 3%2–6% n=6 |
| Clio | n=0 | n=0 | n=0 | n=0 | n=0 | 1%0–3% n=2 | n=0 | n=0 | n=0 | 1%1–2% n=6 | 1%0–2% n=5 | 4%3–5% n=19 | 1%0–1% n=6 | 2%1–3% n=17 | 3%2–4% n=21 | 1%0–3% n=2 |
| Perplexity | n=0 | n=0 | n=0 | n=0 | 1%0–3% n=1 | 1%0–2% n=1 | 1%0–2% n=1 | 2%1–4% n=4 | 1%0–3% n=3 | 0%0–1% n=2 | 1%0–2% n=5 | 0%0–1% n=2 | 1%0–2% n=7 | n=0 | 1%1–2% n=10 | n=0 |
| vLex Vincent | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | 1%0–2% n=1 | 1%0–2% n=2 | 1%0–3% n=3 | 1%0–2% n=5 | 0%0–1% n=2 | 1%1–3% n=7 | 1%0–1% n=5 | 1%0–1% n=5 | n=0 | 1%0–2% n=1 |
| Paxton | n=0 | n=0 | n=0 | 1%0–4% n=1 | 1%0–3% n=1 | 1%0–2% n=1 | 1%0–3% n=2 | 0%0–2% n=1 | 1%0–3% n=3 | n=0 | n=0 | 0%0–1% n=1 | n=0 | 0%0–1% n=1 | 0%0–0% n=1 | n=0 |
| Luminance | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | n=0 | 4%2–6% n=8 | n=0 | n=0 | n=0 | 0%0–1% n=1 | n=0 | n=0 | n=0 | n=0 |
| All named-tool voices | 164 | 269 | 99 | 94 | 146 | 165 | 193 | 228 | 234 | 496 | 603 | 495 | 746 | 793 | 789 | 191 |
2026 Q4*: quarter to date. Read share and net together: a tool can be discussed more because it is used more, not because it is liked less.
By model, by quarter
Posts naming the model in their own words, as a share of that quarter’s posts naming any model of the same family (so GPT-4o fading and GPT-5 arriving shows as a hand-over, not as noise). A net appears only from 5 posts whose main subject is that family.
| Model | Family | 2023 Q1 | 2023 Q2 | 2023 Q3 | 2023 Q4 | 2024 Q1 | 2024 Q2 | 2024 Q3 | 2024 Q4 | 2025 Q1 | 2025 Q2 | 2025 Q3 | 2025 Q4 | 2026 Q1 | 2026 Q2 | 2026 Q3 | 2026 Q4* |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-3.5 | ChatGPT | 111% of ChatGPT model posts | 311% of ChatGPT model posts | 225% of ChatGPT model posts | 17% of ChatGPT model posts | ||||||||||||
| GPT-4o | ChatGPT | 947% of ChatGPT model posts net +80 (+18 to +92, n=5) | 4100% of ChatGPT model posts | 133% of ChatGPT model posts | 450% of ChatGPT model posts | 733% of ChatGPT model posts net +20 (−32 to +60, n=5) | 414% of ChatGPT model posts | 321% of ChatGPT model posts | 19% of ChatGPT model posts | 429% of ChatGPT model posts | 314% of ChatGPT model posts | ||||||
| GPT-4 | ChatGPT | 889% of ChatGPT model posts net +12 (−31 to +50, n=8) | 2796% of ChatGPT model posts net +9 (−20 to +35, n=23) | 788% of ChatGPT model posts net +20 (−34 to +62, n=5) | 7100% of ChatGPT model posts net +17 (−30 to +54, n=6) | 15100% of ChatGPT model posts net +22 (−27 to +62, n=9) | 1263% of ChatGPT model posts net +25 (−16 to +56, n=8) | 524% of ChatGPT model posts | 311% of ChatGPT model posts | 429% of ChatGPT model posts | 327% of ChatGPT model posts | 321% of ChatGPT model posts | 15% of ChatGPT model posts | ||||
| GPT-5 | ChatGPT | 210% of ChatGPT model posts | 1346% of ChatGPT model posts net 0 (−49 to +49, n=7) | 1071% of ChatGPT model posts net +50 (−10 to +81, n=6) | 764% of ChatGPT model posts | 643% of ChatGPT model posts | 1990% of ChatGPT model posts net −11 (−45 to +27, n=9) | ||||||||||
| o1 | ChatGPT | 267% of ChatGPT model posts | 450% of ChatGPT model posts | 314% of ChatGPT model posts | 19% of ChatGPT model posts | ||||||||||||
| o3 | ChatGPT | 838% of ChatGPT model posts net +67 (+2 to +90, n=6) | 1243% of ChatGPT model posts net +100 (+35 to +99, n=5) | ||||||||||||||
| Claude 3.5 Sonnet | Claude | 5100% of Claude model posts | 2100% of Claude model posts | 150% of Claude model posts | 350% of Claude model posts | 17% of Claude model posts | |||||||||||
| Claude 3.7 Sonnet | Claude | 150% of Claude model posts | 267% of Claude model posts | ||||||||||||||
| Claude Sonnet 4.x | Claude | 114% of Claude model posts | 350% of Claude model posts | 111% of Claude model posts | 427% of Claude model posts | 18% of Claude model posts | |||||||||||
| Claude Opus 4.x | Claude | 457% of Claude model posts | 778% of Claude model posts | 1067% of Claude model posts net +33 (−18 to +66, n=6) | 542% of Claude model posts | 1100% of Claude model posts | |||||||||||
| Claude Opus (unversioned) | Claude | 1100% of Claude model posts | 133% of Claude model posts | 229% of Claude model posts | 117% of Claude model posts | 17% of Claude model posts | 542% of Claude model posts | ||||||||||
| Bard | Gemini | 8100% of Gemini model posts | 2100% of Gemini model posts | 7100% of Gemini model posts | 480% of Gemini model posts | 150% of Gemini model posts | 133% of Gemini model posts | 133% of Gemini model posts | 316% of Gemini model posts | ||||||||
| Gemini 2.x | Gemini | 267% of Gemini model posts | 1368% of Gemini model posts net +56 (+8 to +79, n=9) | 686% of Gemini model posts | 433% of Gemini model posts | 343% of Gemini model posts | 125% of Gemini model posts | ||||||||||
| Gemini 3 | Gemini | 343% of Gemini model posts | 650% of Gemini model posts | 8100% of Gemini model posts | 343% of Gemini model posts | 250% of Gemini model posts | |||||||||||
| Gemini Advanced / Pro | Gemini | 120% of Gemini model posts | 1100% of Gemini model posts | 267% of Gemini model posts | 316% of Gemini model posts | 114% of Gemini model posts | 217% of Gemini model posts | 114% of Gemini model posts | 125% of Gemini model posts | ||||||||
| Lexis+ AI Protégé | Lexis+ AI | 6100% of Lexis+ AI model posts net 0 (−53 to +52, n=5) | 5100% of Lexis+ AI model posts | 6100% of Lexis+ AI model posts net −20 (−61 to +34, n=5) | 13100% of Lexis+ AI model posts net +57 (+2 to +82, n=7) | 12100% of Lexis+ AI model posts net +14 (−40 to +62, n=7) | 28100% of Lexis+ AI model posts net +24 (−7 to +48, n=21) | 22100% of Lexis+ AI model posts net −44 (−66 to −9, n=16) | 2100% of Lexis+ AI model posts | ||||||||
| Westlaw Precision | Westlaw AI | 5100% of Westlaw AI model posts net −60 (−81 to −5, n=5) | 2100% of Westlaw AI model posts | 5100% of Westlaw AI model posts | 133% of Westlaw AI model posts | 788% of Westlaw AI model posts | 3100% of Westlaw AI model posts | 4100% of Westlaw AI model posts | 250% of Westlaw AI model posts | 375% of Westlaw AI model posts | |||||||
| Westlaw Edge | Westlaw AI | 1100% of Westlaw AI model posts | 267% of Westlaw AI model posts | 112% of Westlaw AI model posts | 250% of Westlaw AI model posts | 125% of Westlaw AI model posts | |||||||||||
| CoCounsel (Casetext era) | CoCounsel | 1100% of CoCounsel model posts | 16100% of CoCounsel model posts net +62 (+23 to +80, n=13) | 3100% of CoCounsel model posts | 5100% of CoCounsel model posts | 13100% of CoCounsel model posts net −10 (−51 to +36, n=10) | 12100% of CoCounsel model posts net +22 (−15 to +51, n=9) | 10100% of CoCounsel model posts net +50 (+8 to +74, n=10) | 3100% of CoCounsel model posts | 15100% of CoCounsel model posts net +29 (−18 to +61, n=7) | 16100% of CoCounsel model posts net 0 (−45 to +44, n=8) | 4100% of CoCounsel model posts | 6100% of CoCounsel model posts | 6100% of CoCounsel model posts | 6100% of CoCounsel model posts | 1100% of CoCounsel model posts |
Events, with sources
Dated launches, changes, studies and court orders from the vendors’ and providers’ own pages, or the court’s. 44 events, read on Oct 22, 2026. They are context for the marked changes above; the data does not prove that any event caused any shift.
| Date | Kind | About | Event | Source |
|---|---|---|---|---|
| Nov 30, 2022 | Model launch | ChatGPT | OpenAI introduces ChatGPT | archive copy, captured May 2, 2024 |
| Feb 6, 2023 | Model launch | Gemini · Bard | Google announces Bard experimental conversational AI | Google blog: Bard and AI search updates |
| May 4, 2023 | Product launch | Lexis+ AI | LexisNexis announces Lexis+ AI commercial preview | LexisNexis press room: Lexis+ AI commercial preview |
| Jun 26, 2023 | Acquisition | CoCounsel | Thomson Reuters signs agreement to acquire Casetext | Thomson Reuters press release |
| Jul 11, 2023 | Model launch | Claude · Claude 2 | Anthropic announces Claude 2 | Anthropic: Claude 2 |
| Sep 21, 2023 | Product launch | Copilot | Microsoft announces Copilot for Microsoft 365 availability date and price | Microsoft 365 blog: Copilot general availability |
| Nov 13, 2023 | Product launch | Westlaw AI | Thomson Reuters launches AI-Assisted Research on Westlaw | Thomson Reuters: AI-Assisted Research on Westlaw |
| Nov 15, 2023 | Product launch | Copilot | Microsoft states Copilot for Microsoft 365 generally available since Nov 1 | Microsoft 365 blog: Introducing Microsoft Copilot |
| Nov 16, 2023 | Guidance | General | State Bar of California publishes generative AI practical guidance | State Bar of California: Ethics & Technology Resources ( |
| Dec 6, 2023 | Model launch | Gemini · Gemini 1.x | Google introduces Gemini 1.0 | Google blog: Introducing Gemini |
| Dec 19, 2023 | Funding | Harvey | Harvey announces $80M Series B at $715M valuation | Harvey blog: Series B |
| Jan 11, 2024 | Product launch | Lexis+ AI | LexisNexis announces Lexis+ AI commercial preview in Canada and UK | LexisNexis press room: Lexis+ AI preview in Canada and UK |
| Jan 23, 2024 | Guidance | General | Florida Bar news reports Board of Governors adopts generative AI ethics guidelines | The Florida Bar News (dated Jan 23, 2024) |
| Feb 8, 2024 | Model change | Gemini · Gemini 1.x | Google renames Bard to Gemini and launches Gemini Advanced with Ultra 1.0 | Google blog: Bard becomes Gemini |
| Feb 15, 2024 | Model launch | Gemini · Gemini 1.x | Google announces Gemini 1.5 | Google blog: Gemini 1.5 |
| Mar 4, 2024 | Model launch | Claude · Claude 3 Opus | Anthropic announces the Claude 3 model family | Anthropic: Introducing the next generation of Claude |
| May 13, 2024 | Model launch | ChatGPT · GPT-4o | OpenAI announces GPT-4o | archive copy, captured May 16, 2024 |
| May 23, 2024 | Study | General | Stanford RegLab/HAI preprint on hallucination rates of legal AI research tools | Stanford HAI: AI on Trial (published May 23, 2024) |
| Jun 21, 2024 | Model launch | Claude · Claude 3.5 Sonnet | Anthropic launches Claude 3.5 Sonnet | Anthropic: Claude 3.5 Sonnet |
| Jul 29, 2024 | Guidance | General | ABA issues Formal Opinion 512 on generative AI tools | ABA Formal Opinion 512 (copy of the opinion hosted on mcle.o |
| Feb 12, 2025 | Funding | Harvey | Harvey announces $300M Series D led by Sequoia at $3B valuation | Harvey blog: Harvey Raises $300M Series D Led by Sequoia |
| Mar 25, 2025 | Model launch | Gemini · Gemini 2.x | Google announces Gemini 2.5 (Pro Experimental) | Google blog: Gemini 2.5 |
| Apr 14, 2025 | Model launch | ChatGPT · GPT-4.1 | OpenAI launches GPT-4.1, mini and nano in the API | archive copy, captured Apr 15, 2025 |
| May 22, 2025 | Model launch | Claude · Claude Opus 4.x | Anthropic introduces Claude Opus 4 and Claude Sonnet 4 | Anthropic: Introducing Claude 4 |
| Jun 6, 2025 | Court order | General | England and Wales Divisional Court judgment on fake citations in Ayinde and Al-Haroun | National Archives: [2025] EWHC 1383 (Admin) (judgment dated |
| Jun 17, 2025 | Model launch | Gemini · Gemini 2.x | Google expands Gemini 2.5 model family | Google blog: Gemini 2.5 model family expands |
| Jul 22, 2025 | Product launch | Lexis+ AI | LexisNexis launches Protégé in Canada within Lexis+ AI | LexisNexis press room: Protégé in Canada |
| Jul 23, 2025 | Court order | General | N.D. Ala. sanctions order in Johnson v. Dunn over AI-fabricated citations | Johnson v. Dunn, Dkt 204 (Docket Alarm copy of court order) |
| Aug 5, 2025 | Product launch | CoCounsel | Thomson Reuters launches CoCounsel Legal with Deep Research | Thomson Reuters press release |
| Aug 5, 2025 | Model launch | Claude · Claude Opus 4.x | Anthropic releases Claude Opus 4.1 | Anthropic: Claude Opus 4.1 |
| Aug 7, 2025 | Model launch | ChatGPT · GPT-5 | OpenAI introduces GPT-5 | archive copy, captured Aug 10, 2025 |
| Sep 29, 2025 | Model launch | Claude · Claude Sonnet 4.x | Anthropic releases Claude Sonnet 4.5 | Anthropic: Claude Sonnet 4.5 |
| Oct 14, 2025 | Product launch | Westlaw AI | Thomson Reuters announces Deep Research on Practical Law beta | Thomson Reuters press release |
| Nov 10, 2025 | Acquisition | Clio | Clio closes US$1B acquisition of vLex and announces Series G | Clio press release (hosted by JMI) |
| Nov 24, 2025 | Model launch | Claude · Claude Opus 4.x | Anthropic releases Claude Opus 4.5 | Anthropic: Claude Opus 4.5 |
| Dec 10, 2025 | Product launch | Lexis+ AI | LexisNexis unveils Protégé General AI | LexisNexis press room: Protégé General AI |
| Mar 10, 2026 | Funding | Legora | Legora announces $550M Series D at $5.55B valuation | Legora newsroom press release (dateline Mar 10, 2026) |
| Jun 30, 2026 | Model launch | Claude · Claude Sonnet 5.x | Anthropic releases Claude Sonnet 5 | Anthropic: Claude Sonnet 5 |
| Jul 24, 2026 | Model launch | Claude · Claude Opus 5.x | Anthropic releases Claude Opus 5 | Anthropic: Claude Opus 5 |
| Sep 1, 2026 | Product launch | Harvey · Claude Fable 5.1 | Harvey makes Claude Fable 5.1 available in its platform | Harvey blog: Fable 5.1, Now Live in Harvey |
| Sep 9, 2026 | Funding | Harvey | Harvey announces $550M round at $15.5B valuation | Harvey blog |
| Sep 22, 2026 | Model launch | Claude · Claude Opus 5.x | Anthropic introduces Claude Opus 5.5 | Anthropic: Claude Opus 5.5 |
| Sep 28, 2026 | Model launch | Claude · Claude Sonnet 5.x | Anthropic introduces Claude Sonnet 5.5 | Anthropic: Claude Sonnet 5.5 |
| Oct 7, 2026 | Model launch | Claude · Claude Haiku 5.5 | Anthropic introduces Claude Haiku 5.5 | Anthropic: Claude Haiku 5.5 |