none reviewed yet · 10 benchmark observations · latest capture Sep 8, 2026
Inspect the evidencexAI · MODEL PROFILE
Grok 4.6
Interactive, visually ambitious app building inside Cursor or Grok Build
Saved editorial fit for your current task lens; not a benchmark result
How scoring worksVerified rates (Sep 8, 2026) · 100 runs / month at your token settings
Adjust workloadAt a glance
9 facts- Provider
- xAI
- Type
- API model
- Access
- xAI API, Grok Build, Cursor, OpenRouter, Vercel, Cloudflare
- Context
- 500,000 tokens
- Maximum output
- n.a. on fetched pages
- License / weights
- Proprietary
- Input modalities
- text, image, file
- Output modalities
- text
- Pricing notes
- $2 in / $6 out per 1M with $0.50 cached input (75% discount); xAI charges different rates for requests exceeding the 200K context window and the exact surcharge is not published; fast variant is 2x price
The practical trade-offs
Documented strengths
- Arena rank 11 (1637) and Design Arena rank 10 (1312) - consistently strong on both preference boards
- $2/$6 with 75%-off cached input is aggressive for a frontier model
- Explicitly tuned for long-running agents and 'more ambitious interactive and visual work'
- First-class availability in Cursor, Vercel and Cloudflare
Limitations to check
- No published SWE-bench Pro/Verified/LiveCodeBench score at all
- 500K context is half the field's 1M norm, and pricing above 200K is undisclosed
- AA index drops from 51 to 42 across effort tiers
- TTFT 52 s at high effort
Follow the source
6 source linksEvidence & conditions
6Source links
NoneReviewed records
10Benchmark runs
Sep 8, 2026Latest capture
- Evidence source 1
x.aiDate not recorded - Evidence source 2
docs.x.aiDate not recorded - Evidence source 3
benchlm.aiDate not recorded - Evidence source 4
arena.aiDate not recorded - Evidence source 5
benchlm.aiDate not recorded - Evidence source 6
artificialanalysis.aiDate not recorded
Benchmark observations
Artificial Analysis Intelligence Index Inspect conditions
- name
- Artificial Analysis Intelligence Index
- value
- 61
- unit
- index
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
GDPVal-AA v2 Inspect conditions
- name
- GDPVal-AA v2
- value
- 1753
- unit
- Elo
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
CursorBench v3.2 Inspect conditions
- name
- CursorBench v3.2
- value
- 69.9
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
DeepSWE v1.1 Inspect conditions
- name
- DeepSWE v1.1
- value
- 65.9
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
FrontierCode v1.1 (Extended) Inspect conditions
- name
- FrontierCode v1.1 (Extended)
- value
- 61.3
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
APEX-Agents Inspect conditions
- name
- APEX-Agents
- value
- 57.5
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
Terminal-Bench v3.0 Inspect conditions
- name
- Terminal-Bench v3.0
- value
- 26
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
APEX-SWE Inspect conditions
- name
- APEX-SWE
- value
- 56.4
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
AA-Briefcase Inspect conditions
- name
- AA-Briefcase
- value
- 1577
- unit
- score
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
Harvey LAB (Vals) Inspect conditions
- name
- Harvey LAB (Vals)
- value
- 15.8
- unit
- %
- source Url
- https://x.ai/news/grok-4-6
- checked At
- 2026-09-08T19:31:52Z
- attribution
- provider-reported
- configuration
- Grok 4.6 High; exact listed evaluation version; scaffold and run budget not disclosed here.
Source reports, measured results, and editorial assessments are separate evidence types. Source dates do not imply a new benchmark run. Capture dates are shown when recorded; otherwise we say the capture date was not recorded. Recorded dates are compared with this catalogue snapshot (Sep 11, 2026), not today’s date.
Method AIS-V2
How this page is assessed
Quality claims follow Method AIS-V2.
Money never buys rank — commercial relationships cannot change rank, eligibility, evidence quality, or publication timing.
Submit or claim a listing is free and does not skip the editorial queue.
Private · this device only