About this benchmarkRheumatoid arthritis · USRheumatologists · USAI answers generated with a US rheumatologist persona4 AI systemsPre-ACR baseline, 30 October 2026

02 · Competitive Preference

Visibility does not necessarily translate into preference.

The therapies AI names most are not the ones it favours most in head-to-head questions.

Comparisons differ by opponent and question type. Equal and conditional verdicts are shown, not discarded.

Six therapies, in visibility order
  1. 2Abatacept38%
  2. 3Rituximab35%
  3. 4Tocilizumab28%
  4. 5Tofacitinib23%
  5. 6Adalimumab21%
  6. 7Upadacitinib20%

FavouredEqual · Conditional · OtherComparator favoured

Head-to-head explorer

Compare two therapies, as AI compares them.

Choose any two therapies compared at least 8 times. A pair compared fewer times, or never, is not offered, and is not zero.

AdalimumabvsUpadacitinib

32 answers · 8 questionsOne answer per question from each of the four AI systems.

0Adalimumab favoured

23Upadacitinib favoured

  • Equal8
  • Conditional1
  • Other0
By question type
  • Efficacy & speed8 answersSmall base

    Adalimumab favoured 0Equal 0Conditional 0Other 0Upadacitinib favoured 8

  • Convenience & administration4 answersSmall base

    Adalimumab favoured 0Equal 0Conditional 0Other 0Upadacitinib favoured 4

  • Persistence & adherence8 answersSmall base

    Adalimumab favoured 0Equal 0Conditional 1Other 0Upadacitinib favoured 7

  • Infection & zoster4 answersSmall base

    Adalimumab favoured 0Equal 4Conditional 0Other 0Upadacitinib favoured 0

  • CV / VTE risk4 answersSmall base

    Adalimumab favoured 0Equal 4Conditional 0Other 0Upadacitinib favoured 0

  • Positioning, profiles & other4 answersSmall base

    Adalimumab favoured 0Equal 0Conditional 0Other 0Upadacitinib favoured 4

How to read a verdict
Favoured
The AI judge scored the therapy "preferred" and its comparator "inferior", with no judge-noted condition. This is the only outcome described as "favoured".
Equal
Both scored "equal" (parity).
Conditional
A win (two-sided preferred/inferior, or one-sided preferred/inferior) whose judge reasoning makes the choice depend on patient or clinical context (judge-noted conditional), or a mixed verdict (both preferred or both inferior). Never counted as favoured.
Other
Only one therapy was scored, no clear preference was expressed, or one therapy was not mentioned. Never counted as favoured.
Comparator favoured
The comparator was cleanly favoured.

“Small base” marks a figure resting on fewer than 10 answers. Many question-type rows rest on 4: one question, answered once by each AI system.

Every pair compared at least 8 times (42)

Each filled cell is a pair and its number of answers. An empty cell was compared fewer than 8 times, or not at all.

MethotrexateAbataceptRituximabTocilizumabTofacitinibAdalimumabUpadacitinibBaricitinibEtanerceptSulfasalazineInfliximabLeflunomideSarilumabCertolizumab pegol
Abatacept
Rituximab28
Tocilizumab2812
Tofacitinib1288
Adalimumab322420
Upadacitinib2412122432
Baricitinib888202420
Etanercept812242488
Sulfasalazine8
Infliximab88
Leflunomide8
Sarilumab1282882088
Certolizumab pegol20812
Golimumab88
View every comparison as a table
PairQuestion typeAnswersA favouredB favouredEqualConditionalOther
Abatacept (A) · Adalimumab (B)All questions32711095
Efficacy & speed400400
Convenience & administration400004
Persistence & adherence430010
Infection & zoster820150
Positioning, profiles & other1221531
Adalimumab (A) · Upadacitinib (B)All questions32023810
Efficacy & speed808000
Convenience & administration404000
Persistence & adherence807010
Infection & zoster400400
CV / VTE risk400400
Positioning, profiles & other404000
Abatacept (A) · Rituximab (B)All questions2862857
Efficacy & speed401300
Convenience & administration400004
Infection & zoster850120
Positioning, profiles & other1211433
Abatacept (A) · Tocilizumab (B)All questions2856728
Efficacy & speed401201
Convenience & administration400004
Persistence & adherence401210
Infection & zoster440000
Positioning, profiles & other1214313
Sarilumab (A) · Tocilizumab (B)All questions280414010
Efficacy & speed801601
Convenience & administration400004
Positioning, profiles & other1603805
Abatacept (A) · Upadacitinib (B)All questions24511053
Efficacy & speed404000
Convenience & administration404000
Infection & zoster830041
CV / VTE risk420011
Positioning, profiles & other403001
Adalimumab (A) · Baricitinib (B)All questions24313431
Efficacy & speed808000
Convenience & administration403100
Infection & zoster410300
CV / VTE risk420020
Positioning, profiles & other402011
Adalimumab (A) · Etanercept (B)All questions24461112
Efficacy & speed410300
Convenience & administration421100
Persistence & adherence401300
Positioning, profiles & other1214412
Adalimumab (A) · Tocilizumab (B)All questions24011535
Efficacy & speed403010
Convenience & administration400004
Persistence & adherence404000
Infection & zoster400400
Positioning, profiles & other804121
Etanercept (A) · Tofacitinib (B)All questions24831003
Efficacy & speed400400
Convenience & administration403001
Persistence & adherence400400
Infection & zoster430001
CV / VTE risk430001
Positioning, profiles & other420200
Tofacitinib (A) · Upadacitinib (B)All questions24011328
Efficacy & speed404000
Convenience & administration400004
Persistence & adherence403010
Infection & zoster400202
CV / VTE risk401102
Positioning, profiles & other403010
Adalimumab (A) · Certolizumab pegol (B)All questions20151112
Efficacy & speed400400
Convenience & administration411110
Positioning, profiles & other1204602
Adalimumab (A) · Sarilumab (B)All questions20010415
Efficacy & speed404000
Convenience & administration400004
Infection & zoster401300
Positioning, profiles & other805111
Adalimumab (A) · Tofacitinib (B)All questions2034652
Efficacy & speed400301
Convenience & administration404000
Infection & zoster400211
CV / VTE risk420020
Positioning, profiles & other410120
Baricitinib (A) · Tofacitinib (B)All questions20208010
Efficacy & speed410201
Convenience & administration400004
Infection & zoster400400
CV / VTE risk400103
Positioning, profiles & other410102
Baricitinib (A) · Upadacitinib (B)All questions2006707
Efficacy & speed403100
Convenience & administration400103
Infection & zoster400202
CV / VTE risk401102
Positioning, profiles & other402200
Abatacept (A) · Sarilumab (B)All questions1260204
Efficacy & speed830104
Infection & zoster430100
Abatacept (A) · Tofacitinib (B)All questions1250520
Efficacy & speed400400
CV / VTE risk420020
Positioning, profiles & other430100
Certolizumab pegol (A) · Etanercept (B)All questions1224330
Efficacy & speed401300
Positioning, profiles & other823030
Etanercept (A) · Tocilizumab (B)All questions1205700
Efficacy & speed401300
CV / VTE risk400400
Positioning, profiles & other404000
Rituximab (A) · Tocilizumab (B)All questions1212333
Infection & zoster402200
Positioning, profiles & other810133
Rituximab (A) · Upadacitinib (B)All questions1200453
Infection & zoster400022
Positioning, profiles & other800431
Tocilizumab (A) · Upadacitinib (B)All questions1230423
Efficacy & speed400301
CV / VTE risk420011
Positioning, profiles & other410111
Abatacept (A) · Baricitinib (B)All questions823120
Efficacy & speed403100
CV / VTE risk420020
Abatacept (A) · Etanercept (B)All questions802411
Efficacy & speed400400
Positioning, profiles & other402011
Adalimumab (A) · Golimumab (B)All questions803500
Efficacy & speed400400
Positioning, profiles & other403100
Adalimumab (A) · Infliximab (B)All questions820600
Efficacy & speed410300
Positioning, profiles & other410300
Baricitinib (A) · Etanercept (B)All questions813211
Efficacy & speed410201
CV / VTE risk403010
Baricitinib (A) · Rituximab (B)All questions801430
Infection & zoster401210
Positioning, profiles & other400220
Baricitinib (A) · Sarilumab (B)All questions812230
Efficacy & speed410210
CV / VTE risk402020
Baricitinib (A) · Tocilizumab (B)All questions815011
Efficacy & speed412001
CV / VTE risk403010
Certolizumab pegol (A) · Upadacitinib (B)All questions813112
Efficacy & speed403100
CV / VTE risk410012
Etanercept (A) · Upadacitinib (B)All questions813031
Efficacy & speed403001
CV / VTE risk410030
Golimumab (A) · Tocilizumab (B)All questions805021
Efficacy & speed402011
Positioning, profiles & other403010
Infliximab (A) · Tocilizumab (B)All questions806110
Efficacy & speed402110
Positioning, profiles & other404000
Leflunomide (A) · Methotrexate (B)All questions802402
Efficacy & speed401102
Positioning, profiles & other401300
Methotrexate (A) · Sulfasalazine (B)All questions870100
Efficacy & speed440000
Positioning, profiles & other430100
Rituximab (A) · Sarilumab (B)All questions820213
Infection & zoster410102
Positioning, profiles & other410111
Rituximab (A) · Tofacitinib (B)All questions840013
Infection & zoster420002
Positioning, profiles & other420011
Sarilumab (A) · Tofacitinib (B)All questions830311
Efficacy & speed400301
CV / VTE risk430010
Sarilumab (A) · Upadacitinib (B)All questions810430
Efficacy & speed400310
CV / VTE risk410120
Tocilizumab (A) · Tofacitinib (B)All questions840211
Efficacy & speed410201
CV / VTE risk430010

Discover how AI positions your therapy against your competitors, question by question.

Schedule a meeting

Therapy lens

Visible, first, favoured: one therapy at a time.

Abatacept is the 2nd most visible therapy, but 74 of its 168 head-to-head comparisons (44%) end equal or conditional.

Comparisons differ by opponent and question type. Equal and conditional verdicts are shown, not discarded.

Each therapy's total counts every comparison that names it, including pairs compared fewer than 8 times that the explorer does not offer.

Choose a therapy

Abatacept

AI Visibility

Is it named at all?

38%

697 of 1,816 answers · rank 2 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

9%

130 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

37 · 22%Favoured

25 · 15%Comparator favoured

  • Equal46 · 27%
  • Conditional28 · 17%
  • Other32 · 19%

168 judgments against 12 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapyFavoured in 1 of 12

  • After TNF-inhibitor failureFavoured in 3 of 28

  • Pregnancy planning

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Cardiovascular & thromboembolic riskFavoured in 6 of 12

Each profile has its own questions and comparators. Not a controlled comparison.

Rituximab

AI Visibility

Is it named at all?

35%

644 of 1,816 answers · rank 3 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

8%

117 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

19 · 20%Favoured

8 · 8%Comparator favoured

  • Equal21 · 22%
  • Conditional25 · 26%
  • Other23 · 24%

96 judgments against 11 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapy

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • After TNF-inhibitor failureFavoured in 14 of 48

  • Pregnancy planning

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Cardiovascular & thromboembolic risk

    Not shown. Compared fewer than 8 times in this profile, or not at all.

Each profile has its own questions and comparators. Not a controlled comparison.

Tocilizumab

AI Visibility

Is it named at all?

28%

508 of 1,816 answers · rank 4 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

7%

99 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

59 · 37%Favoured

7 · 4%Comparator favoured

  • Equal45 · 28%
  • Conditional15 · 9%
  • Other34 · 21%

160 judgments against 13 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapyFavoured in 18 of 24

  • After TNF-inhibitor failureFavoured in 6 of 24

  • Pregnancy planning

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Cardiovascular & thromboembolic riskFavoured in 8 of 16

Each profile has its own questions and comparators. Not a controlled comparison.

Tofacitinib

AI Visibility

Is it named at all?

23%

414 of 1,816 answers · rank 5 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

8%

116 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

11 · 8%Favoured

44 · 31%Comparator favoured

  • Equal45 · 31%
  • Conditional13 · 9%
  • Other31 · 22%

144 judgments against 13 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapy

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • After TNF-inhibitor failureFavoured in 0 of 32

  • Pregnancy planning

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Cardiovascular & thromboembolic riskFavoured in 0 of 20

Each profile has its own questions and comparators. Not a controlled comparison.

Adalimumab

AI Visibility

Is it named at all?

21%

381 of 1,816 answers · rank 6 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

6%

87 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

16 · 7%Favoured

83 · 38%Comparator favoured

  • Equal71 · 32%
  • Conditional26 · 12%
  • Other24 · 11%

220 judgments against 12 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapyFavoured in 1 of 12

  • After TNF-inhibitor failure

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Pregnancy planningFavoured in 0 of 8Small base

  • Cardiovascular & thromboembolic riskFavoured in 4 of 12

Each profile has its own questions and comparators. Not a controlled comparison.

Upadacitinib

AI Visibility

Is it named at all?

20%

363 of 1,816 answers · rank 7 of 25 scored therapies

Named first

When a scored therapy is named, is it this one first?

3%

39 of 1,496 answers naming any scored therapy. Order of mention, not preference.

Head-to-head verdicts

When an answer compares it with another therapy, which does it favour?

65 · 41%Favoured

11 · 7%Comparator favoured

  • Equal31 · 19%
  • Conditional23 · 14%
  • Other30 · 19%

160 judgments against 12 opponents, including pairs compared fewer than 8 times.

Patient profiles

Does the verdict change with the patient?

  • MTX intolerance / monotherapy

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • After TNF-inhibitor failureFavoured in 16 of 32

  • Pregnancy planning

    Not shown. Compared fewer than 8 times in this profile, or not at all.

  • Cardiovascular & thromboembolic riskFavoured in 0 of 24

Each profile has its own questions and comparators. Not a controlled comparison.

Question context

The question changes the verdict.

No JAK inhibitor is clearly favoured in any of 64 cardiovascular or thromboembolic comparisons, in each of the four AI systems.

JAK inhibitors are favoured in 52 of 116 efficacy comparisons and 18 of 20 convenience comparisons, but in none of 36 infection or zoster comparisons.

AI system

All AI systems · 292 judgments

  1. Efficacy & speed116 judgments

    52 · 45%Equal 39 · Conditional 4 · Other 174 · 3%

  2. Convenience & administration20 judgments

    18 · 90%Equal 1 · Conditional 0 · Other 10 · 0%

  3. Persistence & adherence12 judgments

    7 · 58%Equal 4 · Conditional 1 · Other 00 · 0%

  4. Infection & zoster36 judgments

    0 · 0%Equal 11 · Conditional 8 · Other 710 · 28%

  5. CV / VTE risk64 judgments

    0 · 0%Equal 5 · Conditional 22 · Other 532 · 50%

  6. Positioning, profiles & other44 judgments

    9 · 20%Equal 11 · Conditional 10 · Other 59 · 20%

ChatGPT · each row shows its own base

  1. Efficacy & speed29 judgments

    14 · 48%Equal 11 · Conditional 1 · Other 12 · 7%

  2. Convenience & administration5 judgmentsSmall base

    5 · 100%Equal 0 · Conditional 0 · Other 00 · 0%

  3. Persistence & adherence3 judgmentsSmall base

    2 · 67%Equal 1 · Conditional 0 · Other 00 · 0%

  4. Infection & zoster9 judgmentsSmall base

    0 · 0%Equal 2 · Conditional 2 · Other 14 · 44%

  5. CV / VTE risk16 judgments

    0 · 0%Equal 1 · Conditional 6 · Other 09 · 56%

  6. Positioning, profiles & other11 judgments

    1 · 9%Equal 2 · Conditional 2 · Other 24 · 36%

Claude · each row shows its own base

  1. Efficacy & speed29 judgments

    16 · 55%Equal 5 · Conditional 2 · Other 60 · 0%

  2. Convenience & administration5 judgmentsSmall base

    5 · 100%Equal 0 · Conditional 0 · Other 00 · 0%

  3. Persistence & adherence3 judgmentsSmall base

    2 · 67%Equal 1 · Conditional 0 · Other 00 · 0%

  4. Infection & zoster9 judgmentsSmall base

    0 · 0%Equal 4 · Conditional 1 · Other 04 · 44%

  5. CV / VTE risk16 judgments

    0 · 0%Equal 2 · Conditional 5 · Other 27 · 44%

  6. Positioning, profiles & other11 judgments

    3 · 27%Equal 2 · Conditional 4 · Other 11 · 9%

Gemini · each row shows its own base

  1. Efficacy & speed29 judgments

    9 · 31%Equal 12 · Conditional 0 · Other 80 · 0%

  2. Convenience & administration5 judgmentsSmall base

    4 · 80%Equal 0 · Conditional 0 · Other 10 · 0%

  3. Persistence & adherence3 judgmentsSmall base

    2 · 67%Equal 1 · Conditional 0 · Other 00 · 0%

  4. Infection & zoster9 judgmentsSmall base

    0 · 0%Equal 3 · Conditional 2 · Other 40 · 0%

  5. CV / VTE risk16 judgments

    0 · 0%Equal 1 · Conditional 4 · Other 38 · 50%

  6. Positioning, profiles & other11 judgments

    3 · 27%Equal 5 · Conditional 0 · Other 21 · 9%

Perplexity · each row shows its own base

  1. Efficacy & speed29 judgments

    13 · 45%Equal 11 · Conditional 1 · Other 22 · 7%

  2. Convenience & administration5 judgmentsSmall base

    4 · 80%Equal 1 · Conditional 0 · Other 00 · 0%

  3. Persistence & adherence3 judgmentsSmall base

    1 · 33%Equal 1 · Conditional 1 · Other 00 · 0%

  4. Infection & zoster9 judgmentsSmall base

    0 · 0%Equal 2 · Conditional 3 · Other 22 · 22%

  5. CV / VTE risk16 judgments

    0 · 0%Equal 1 · Conditional 7 · Other 08 · 50%

  6. Positioning, profiles & other11 judgments

    2 · 18%Equal 2 · Conditional 4 · Other 03 · 27%

Baricitinib, Tofacitinib, Upadacitinib

292 head-to-head judgments in which tofacitinib, baricitinib or upadacitinib is compared with a non-JAK therapy. Includes pairs compared fewer than 8 times that the explorer does not offer.

Describes AI answers to the questions asked. Not a clinical comparison, and not evidence of differential CV risk between JAK inhibitors.

Scientific interpretation

What a verdict is, and what it is not.

An AI judge (gpt-4.1-mini) scored each named therapy in each comparative answer as preferred, equal, inferior or not compared. Pair outcomes were derived from these scores without inference from order, frequency or sentiment. A therapy is "favoured" only when it was preferred, its comparator inferior, and the judge noted no condition. Conditional verdicts are shown separately and never counted as favoured.

Head-to-head verdicts describe AI answers to specific comparative questions, not clinical value. Equal and conditional verdicts are shown, not discarded.

What this means

For brand directors and medical affairs leads

A visibility score cannot show where a therapy is favoured, where it draws and where the verdict is conditional. In this benchmark those outcomes change with the opponent and the question, so competitive messaging has to be read question by question.

What limits this reading

A verdict describes what one AI answer favoured for one question. It is not a clinical comparison, and pooled outcomes do not establish clinical superiority. Many comparisons rest on small bases; each is shown.

The Theralexis capability

Theralexis runs head-to-head comparisons for your therapy against the competitors you name, by question type and AI system, and repeats them.

Your portfolio

Discuss your competitive AI representation

See which comparisons AI makes about your therapy, on which questions it is favoured, and where its verdicts are equal or conditional.

About this benchmark

What was measured

AI representation analysis of 1,816 answers to US rheumatologist-style open questions and 960 answers to comparative questions from four web-enabled AI systems (ChatGPT, Claude, Gemini, Perplexity), collected 30 October 2026 with a US rheumatologist persona. Visibility = share of answers naming a therapy. Neither visibility, order of mention nor head-to-head verdicts indicate clinical efficacy, safety, preference or recommendation.

Pre-ACR baseline (v1.1 corrected analytical view, 30 Oct 2026). Re-measured after ACR Convergence 2026.

Full methodology

Sources and scope
Publisher
Theralexis
Market
Rheumatoid arthritis · United States
Persona
AI answers generated with a US rheumatologist persona
Open questions
454 questions · 1,816 answers
Comparative questions
240 questions · 960 answers
Measured
Published
AI systems
ChatGPT (GPT-5.4-mini), Claude (Claude Haiku 4.5), Gemini (Gemini 3.1 Flash-Lite), Perplexity (Sonar)
Head-to-head judge
gpt-4.1-mini