Grok 4.7: score and benchmark evidence
Historical release · · xAI
111.3 points · Rank 11 among 25 ranked models · Stability range 94.2–129.6 points.
Rank follows the point estimate; it does not establish statistically significant superiority. The scale is anchored at mean 100, SD 15 in the frozen calibration cohort, not human IQ or a percentage.
One score summarizes demonstrated capability across published reasoning settings and seven equally weighted domains. Related benchmark variants share a family budget, and influence from one evaluator is capped.
Release tli-2026-v1.0-2026-09-22-gpt6pair-1. Methodology and limitations for this release.
This page preserves historical evidence. See the current model record.
Capability domains
Domain scores use calibration-cohort standard-deviation units. Unmeasured domains are shown as unavailable, not zero.
| Domain | Score (SD units) | Observed benchmarks |
|---|---|---|
| Knowledge & reasoning | 0.32 | 3 |
| Coding & software engineering | 0.47 | 3 |
| Agentic tool & computer use | Unavailable | 0 |
| Professional & real-world work | Unavailable | 0 |
| Multimodal & vision | Unavailable | 0 |
| Cybersecurity | Unavailable | 0 |
| Preference & communication | Unavailable | 0 |
Published benchmark evidence
6 results contribute to this score, from 16 collected results and 6 contributing benchmark families. Counts are not independent sample sizes.
Results, source citations, review decisions and compatible reasoning modes are stored in a versioned relational database. Each publication uses a sealed, checksum-verified export; CSV files are optional exports, not editable scoring inputs. Previous revisions remain available for audit. Collected means a numeric result exists in the database. Used means its source, exact model checkpoint and benchmark protocol have been reviewed and its benchmark has a frozen calibration supported by multiple providers. For each benchmark we select the highest published aggregate across comparable reasoning modes, not necessarily maximum effort, and count it once. More tested settings provide more chances to record a high score; this selection advantage is not corrected by the conditional interval. Developer-reported and sole/unspecified settings remain eligible and disclosed. Unresolved does not mean false or useless: conflicting metrics, unverified citations and unmatched versions stay collected until reconciled.
The score stability range is a 95% bootstrap score interval conditional on frozen calibration and comparable evidence availability. It resamples benchmark families and evaluators and adds a frozen missing-evidence correction. It excludes selection across reasoning settings, selective publication, calibration changes and new domains. It is not a 95% guarantee about true capability or where a future score will land. Sparse or conflicting evidence can widen the range.
Reporting sensitivity: ±24.7 points. Sensitivity to selectively published results. This stress range is not a confidence interval or a probability statement.
CursorBench 4.0: 46.3%
Not used in the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (xhigh); CursorBench 4.0 published evaluation; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
DeepSWE v1.1: 71%
Contributes to the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (high); xAI launch evaluation; harness undisclosed; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted. The launch table explicitly marks this result high effort, not xhigh.
EEBench: 64%
Not used in the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (xhigh); xAI launch evaluation; harness undisclosed; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
Terminal-Bench 4.0: 38%
Contributes to the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (xhigh); xAI launch evaluation; harness undisclosed; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
Harvey Legal Agent Benchmark — xAI release: 19.6%
Not used in the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (xhigh); xAI launch evaluation; harness undisclosed; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
HealthBench Professional — xAI release: 56.7%
Not used in the score. Developer-reported. Review status: reviewed-developer.
Configuration: Grok 4.7 (xhigh); xAI launch evaluation; harness undisclosed; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted. Length adjustment is undisclosed; this does not populate the calibrated length-adjusted row.
AA-Briefcase v1.1 (Elo): 1657.2 native-points
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: Elo
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted. Independent 1657.2 xhigh corroborates the rounded 1657 in the supplied launch table.
GDPval-AA v2.1 (Elo): 1695.21 native-points
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: Elo
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
AutomationBench-AA: 65.5583514897%
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
Terminal-Bench 4.0 — AA harness: 25.7575757576%
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
SciCode: 57.7546296296%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (high); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
Humanity's Last Exam (no tools): 43.1417979611%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
GDP.pdf — AA harness: 23.2%
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (high); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
CritPt: 18%
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (high); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
AA-Omniscience: 32.0333333333 native-points
Contributes to the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (xhigh); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: omniscience index
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
AA-LCR v1.1: 77%
Not used in the score. Source category: independent. Review status: reviewed-independent.
Configuration: Grok 4.7 (high); Artificial Analysis independent evaluation, September 21 2026; retrieved 2026-09-21
Metric: published benchmark score (%)
Original source · Retrieved 2026-09-21
Reviewed published aggregate for the exact checkpoint and effort. Retry budget and fallback accounting are undisclosed. Separate benchmark versions and evaluator harnesses are not substituted.
Cite this model record
The Latent. “Grok 4.7: score and benchmark evidence.” 2026-09-22. Release tli-2026-v1.0-2026-09-22-gpt6pair-1.
Download release data and definitions (JSON) · Data usage terms