Skip to content

xAI Proprietary Released Aug 12, 2026 Updated

Grok 4.6

ToolsGuide Scores, rankings, test results and price.

Model summary

  • Text #71 / 416

    433 / 1,000

    ToolsGuide Score· votes only

  • Vision #45 / 167

    476 / 1,000

    ToolsGuide Score· votes only

  • Coding #29 / 147

    458 / 1,000

    ToolsGuide Score· votes 463, tests 452

  • Agents #27 / 53

    471 / 1,000

    ToolsGuide Score· votes only

  • Price per 1M tokens

    In $2.00· Out $6.00

    $2.00

    Input · median text model $0.43

Summary

Grok 4.6 is xAI's model, released Aug 12, 2026. It's ranked in 4 categories and does best at text: #71 of 416, with a ToolsGuide Score of 433 against a median of 235.

It also ranks #29 of 147 for coding (458), #45 of 167 for vision (476), #27 of 53 for agents (471).

Its text and vision and agents score uses votes only, because it has published fewer than two of the tests used for those categories.

It costs $2.00 per million input tokens and $6.00 per million output tokens, about 4.6× the median text model ($0.43).

It has published results on 11 tests and has the top score on 3 of them, including AA-Briefcase and FrontierCode.

Specifications

Creator
xAI
Licence
Proprietary
Released
Aug 12, 2026
Context window
500K tokens
Input price
$2.00 / 1M
Output price
$6.00 / 1M
Tests reported
11
Settings ranked
grok-4.6-high, Grok 4.6 (High), Grok 4.6 (xHigh)

Model comparison

Against the ten best text models.

Text score

ToolsGuide Score out of 1,000 · Higher is better

Coding score

ToolsGuide Score out of 1,000 · Higher is better

Input price

USD per 1M input tokens · Lower is better

Leaderboards

Where Grok 4.6 sits among the top 20 in each category.

Text

#71 of 416 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full text leaderboard →

Vision

#45 of 167 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full vision leaderboard →

Coding

#29 of 147 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full coding leaderboard →

Agents

#27 of 53 · ToolsGuide Score out of 1,000, where 1,000 is a flawless model

Full agents leaderboard →

By task

Scored from votes on each kind of prompt alone, on the same 1,000-point scale.

Score against price

Text · the 10 priced models ranked closest · up and to the left is better value

Against its nearest rivals

Text prompts by kind · ToolsGuide Score from votes · hover for values

Text

grok-4.6-high · hover a bar for its rank

Vision

Grok 4.6 (High) · hover a bar for its rank

Coding

Grok 4.6 (High) · hover a bar for its rank

Test results

Every test Grok 4.6 has a published result on, out of 100%, with its rank among the models that took it.

  • DeepSWE 1.1

    #29 / 50

    DeepSWE 1.1 evaluates software engineering agents using the mini-swe-agent harness.

    65.9%

    Self-reported · In the coding score

  • Terminal-Bench 4.0

    #23 / 28

    Terminal-Bench 4.0 evaluates AI agents on real-world work in terminal and command-line containerized environments across 66 tasks, with emphasis on science-adjacent and frontier engineering problems. Relative to earlier Terminal-Bench releases, 4.0 increases timeouts and adaptively raises RAM/CPU on selected tasks to reduce harness and resource confounds.

    20.3%

    Self-reported · In the coding score· In the agents score

  • GDPval-AA

    #5 / 14

    GDPval-AA evaluates AI agents on economically valuable professional knowledge-work tasks and reports performance as an Elo score.

    58.4%

    Self-reported

  • APEX-Agents

    #2 / 11

    APEX-Agents is a benchmark evaluating AI agents on long horizon professional tasks that require sustained reasoning, planning, and execution across complex multi-step workflows.

    57.5%

    Self-reported

  • AA-Briefcase

    #1 / 4

    AA-Briefcase is an Artificial Analysis evaluation of AI systems on professional knowledge-work tasks, reported as an Elo score.

    52.6%

    Self-reported

  • AA-Index

    #3 / 4

    No official academic documentation found for this benchmark. Extensive research through ArXiv, IEEE/ACL/NeurIPS papers, and university research sites yielded no peer-reviewed sources for an 'aa-index' benchmark. This entry requires verification from official academic sources.

    61.0%

    Self-reported

  • FrontierCode

    #1 / 4

    FrontierCode is Cognition's coding evaluation that tests whether models can pass difficult coding tasks while meeting the standards of high-quality production codebases. The Diamond subset contains the hardest problems.

    61.3%

    Self-reported

  • Terminal-Bench 3.0

    #3 / 4

    Terminal-Bench 3.0 is a release of the Terminal-Bench benchmark that tests AI agents' ability to operate a computer via the terminal on real-world, end-to-end tasks.

    26.0%

    Self-reported

  • CursorBench v3.2

    #2 / 2

    CursorBench v3.2 evaluates coding agents on interactive software engineering tasks in the Cursor environment.

    69.9%

    Unverified

  • Harvey LAB (Vals)

    #2 / 2

    Harvey LAB (Vals) is a professional legal-work evaluation of AI systems on complex law-firm style tasks, reported by Vals.

    15.8%

    Unverified

  • APEX-SWE

    #1 / 1

    APEX-SWE evaluates AI agents on software engineering tasks requiring multi-step coding, debugging, and verification.

    56.4%

    Unverified