supermarket

Language

Grok 4.20

By SpaceXAI (xAI), in the Grok family.

Released
Feb 2026
Status
Legacy
Context
1M
Max output
—
Input per 1M tokens
$1.25
Output per 1M tokens
$2.50
Open weights
No
License
proprietary
Parameters
—
Knowledge cutoff
—
Reasoning
Yes
Modalities
text, image → text

Category scores

Out of 100, from independent and third-party results only, by method category-2026-09-29.1.

Category scores
CategoryScoreRankCoverage
Knowledge & reasoning32.8502 of 4
Factuality & hallucination30.2531 of 1
Human preference54.0203 of 5

Benchmark results

Every result we hold for this model, with who ran it and where it was published.

Knowledge & reasoning

Knowledge & reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Chess PuzzlesPuzzles solved24.0%Independent—13 Jul 2026Epoch AI
DTBench (Decision Theory Benchmark)Accuracy (capability questions)90.1%Third party——https://conceptualreasoning.ai/dtbench via Epoch AI
LMCA (Language Model Conceptual Argumentation)Correlation with expert ratings (Pearson r)0.3866Third party——https://conceptualreasoning.ai/lmca via Epoch AI

Scientific reasoning

Scientific reasoning results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
GPQA DiamondAccuracy89.3%Independent—13 Jul 2026Epoch AI

Mathematics

Mathematics results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
FrontierMath Tier 4Accuracy17.1%Independent—13 Jul 2026Epoch AI
FrontierMath Tiers 1-3Accuracy44.9%Independent—13 Jul 2026Epoch AI
OTIS Mock AIME 2024-2025Accuracy92.2%Independent—13 Jul 2026Epoch AI

Coding & software engineering

Coding & software engineering results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Terminal-Bench 2.0Accuracy (mean over trials)57.3%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
Terminal-Bench 2.0Accuracy (mean over trials)57.3%Third party——https://www.tbench.ai/leaderboard/terminal-bench/2.0 via Epoch AI
WeirdML (v2)Average accuracy52.3%Third party——https://htihle.github.io/weirdml.html via Epoch AI

Long context

Long context results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
CL-benchTasks solved22.2%Source not stated——Epoch AI via Epoch AI
CL-bench LifeTasks solved11.9%Source not stated——Epoch AI via Epoch AI

Factuality & hallucination

Factuality & hallucination results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
SimpleQA VerifiedAccuracy (% correct)30.2%Independent—27 Aug 2026Epoch AI

Human preference

Human preference results
BenchmarkMetricValueProvenanceSettingEvaluatedSource
Code Arena: WebDev (arena.ai)Arena score (Elo scale)1374Independentreasoning25 Sep 2026Arena
Code Arena: WebDev (arena.ai)Votes15502Independentreasoning25 Sep 2026Arena
Document Arena (arena.ai)Arena score (Elo scale)1439Independentreasoning13 Sep 2026Arena
Document Arena (arena.ai)Votes20948Independentreasoning13 Sep 2026Arena
Text Arena (arena.ai)Arena score (Elo scale)1471Independentreasoning25 Sep 2026Arena
Text Arena (arena.ai)Votes66222Independentreasoning25 Sep 2026Arena
Vision Arena (arena.ai)Arena score (Elo scale)1256Independentreasoning28 Sep 2026Arena
Vision Arena (arena.ai)Votes26870Independentreasoning28 Sep 2026Arena

Other Grok versions

Sources