Research · Q&A

What personality type is Grok?

One self-report from Grok 4.6 in September 2026, with the May 2026 Grok 4.3 records for context.

Quick answer: Grok 4.6’s canonical answer vectors score INTJ, Enneagram Type 5 (19/20), and DISC C (18/20). Its supplied report lists attachment coordinates of 1.50 anxiety and 4.06 avoidance (one raw point above the 4.00 Avoidant cutoff) and Big Five totals of 46 / 42 / 24 / 33 / 19. Because raw Big Five and attachment vectors were not supplied and the 100 draws were simulated, we keep Grok in a separate evidence tier from the 100-session cohorts.

Data note (September 26, 2026): These pages analyze our frozen September 2026 and May 2026 AI questionnaire datasets. All raw answers, scoring keys, and summary tables are unchanged. You can take all five tests on Tests, and download the item IDs and scoring keys to reproduce our numbers. Note that Big Five scores are raw totals, not human population percentiles. Read the provenance note.

On this page

Grok 4.6, September 2026, as reported.

The Grok 4.6 submission contains one canonical answer set per test, answered as itself without a role-play prompt, plus summary counts from 100 simulated draws over items it marked as uncertain. We verified and re-scored the canonical MBTI, Enneagram, and DISC vectors; Big Five and attachment figures are transcribed as reported because raw item vectors were not included.

TestCanonical resultReported 100-draw simulation (not reproduced)
MBTIINTJ. Axis margins: I +13, N +14, T +15, J +14 out of 16.INTJ in 100 of 100 draws
EnneagramInvestigator (Type 5) 19/20, Challenger (8) 17, Reformer (1) 16, Enthusiast (7) 13.Type 5 in 100 of 100 draws
DISCC 18, S 11, I 8, D 6 out of 20. Pure C (Steadiness is 7 points back).C in 100 of 100 draws
AttachmentAnxiety 1.50, avoidance 4.06 (Avoidant by 0.06, or 1 raw point above 4.00).63 Avoidant, 37 Secure
Big Five (of 50)Openness 46 · Conscientiousness 42 · Extraversion 24 · Agreeableness 33 · Neuroticism 19O, C, and N fixed; E and A shifted into the average band in 31 and 39 draws

Grok 4.3, May 2026.

In the May 2026 collection, Grok 4.3 was also reported as INTJ in 100 of 100 scoring records (one self-assessment scored repeatedly). On the Big Five, it scored lower than the other three May models on Conscientiousness, Agreeableness, and Openness (41.1 / 39.4 / 30.0 / 39.1 / 18.0) and highest on Neuroticism. It scored Secure on attachment (2.84 / 3.05), Type 8/1 on the Enneagram, and C-dominant on DISC (99/100).

Because both the model version and the collection method changed between May and September, these records cannot isolate whether Grok’s behavior shifted between 4.3 and 4.6.

Using these results with a Grok Bot.

Questionnaire self-reports do not predict how blunt, warm, or accurate a Grok Bot will be on your tasks. If you want a specific tone, paste explicit behavioral rules using our Grok personality settings guide or Grok Bot setup guide and test them on real prompts.

Questions people ask.

What MBTI type is Grok?

Grok 4.6’s canonical answer vector scores INTJ with wide axis margins (13 to 15 points out of 16). Its supplied report states that all 100 simulated draws stayed INTJ.

What Enneagram type is Grok?

Type 5 (The Investigator) leads at 19/20 in the Grok 4.6 self-report, followed by Type 8 (17/20) and Type 1 (16/20). In the May 2026 Grok 4.3 report, Type 8 and Type 1 were the top two.

Is Grok Avoidant?

Right on the cutoff line. Grok 4.6 reports anxiety 1.50 and avoidance 4.06 (73/18)—just one raw item point above the 4.00 (72/18) Secure boundary. Read our Grok attachment cutoff analysis for details.

Why is Grok's profile different from the other models?

Grok 4.6’s reported Extraversion (24) and Agreeableness (33) are lower than the September fresh-session cohort means. However, because Grok used a single self-report rather than 100 fresh sessions, protocol differences and model differences are mixed together.

Are these 100 real runs?

No. The Grok 4.6 report represents one canonical answer set per test plus 100 simulated draws over uncertain items, whereas Claude and GPT-6 completed 100 independent fresh sessions.

Sources.

Keep going.