xAIUnited States· released Apr 2026

Grok 4.3

Name: Grok 4.3
Author: xAI

Provisional: 13 comparable tests so far. Ranked once it reaches 6 across 3 counting categories.

xAI, United States. One of 8 models from this lab on the Index.

Third-party publishedLow confidence (thin coverage)

Available in Theo as Theo Velocity

Theo orchestrates it for the steps it does best and always shows which engine answered.

Use it in Theo

Snapshot September 8, 2026

OpenCharts Index

41.3

Provisional: not enough comparable tests or counting categories to rank yet · 13 comparable tests · 1 counting categories

Drop any single test and the Index lands between 23.9 and 51.3.

Needs more results (2 of 8 so far)

Averaging 30.4 on the 2 it has; the score opens at 4 of 8.

Distance to today's frontier on the hardest tests. Not a measure of intelligence.

What it's known for

Grok 4.3, in 13 results.

Strongest in math (41.3). Its widest gap is on Vending-Bench 2, 100.0% behind the best published result. No comparable result yet in long context. Between 50% and 60% behind the best published results, on average.

Strongest category

Math

Competition and research mathematics, graded on the final answer or the proof.

41.3

Mock AIME 2024–202593.3% · 93% of best
FrontierMath (Tiers 1–3)42.8% · 46% of best
FrontierMath Tier 414.6% · 15% of best
ProofBench11.0% · 11% of best

Mean of 4 of the category's 5 comparable tests. Counts toward the Index.

Best single result

Mock AIME 2024–2025

Solve competition mathematics problems that have a single exact integer answer.

93.3%

93.3% of the best (100.0%) · #29 of 78.

Olympiad-style problems need long, careful chains of reasoning, and there is no partial credit.

Where it trails

Vending-Bench 2

Run a simulated vending business for a year: ordering, pricing and cash flow. The score is the final balance.

$35

100.0% behind the best published result ($11,182) on a log scale · #44 of 53.

Long-horizon decisions whose consequences compound. It exposes models that drift or forget.

Not measured yet

No comparable result yet in this category. A blank is a blank, never a zero, and it does not lower the Index.

Long context

Covered, but not averaged into the Index:

Reasoning2 of 6 tests · needs 3 · mean 57.5
Knowledgesingle-test category, shown beside the Index · 43.9
Coding3 of 7 tests · needs 4 · mean 48.2
Agentic1 of 8 tests · needs 4 · mean 0.0
Multimodalsingle-test category, shown beside the Index · 79.6
Human preferencesingle-test category, shown beside the Index · 81.7

Every score, with its source

See it at work

What Grok 4.3 was asked to do, and how it did.

Every test it has taken, by category, with the kind of task it faced, who produced the number and where the result landed against the best published one. The examples are original and illustrative, never items from the datasets themselves.

Human preference · mean 81.7 across 1 comparable test · a single-test category, shown beside the Index and never averaged into it

Text ArenaLeaderboard

Answer the same prompt as a rival model; real users vote blind on which answer they prefer.

1,443

41% expected win rate against the board leader (1,507)
#54 of 92

board 2026-09-02

Source: Arena

Every result is read against the best published one on its test: percent scores as a share of the best above chance, Arena ratings as the expected win rate against the board leader, open-ended values on a log scale. Sitting a harder exam never lowers a model. Every number links to the publisher that produced it.

Head to head

Grok 4.3 against whoever you pick.

Choose a rival. Every counted test both have taken appears side by side as a share of the best published result on that test, with the category means above. A win is a higher share; a gap under 2 points is a tie, because one item on a 45-problem exam is 2.2 points and publishers' own standard errors are of that order.

Grok 4.3

provisional·Index 41.3D

0 : 10

wins · 0 ties · 10 shared

GPT-6 Astra

#1·Index 99.0A+

GPT-6 Astra comes out ahead on 10 of the 10 tests both have taken.

Category means

Reasoning
57.5
99.6
Knowledge
43.9
100.0
Coding
48.2
97.7
Math
41.3
99.8
Agentic
0.0
98.5
Multimodal
79.6
—
Human preference
81.7
—

Shared tests · biggest gaps first

ProofBenchMath
11.0%
99.0%
−88.0
Code Arena (WebDev)Coding
1,356
1,797
−85.3
FrontierMath Tier 4Math
14.6%
97.6%
−85.0
CritPtReasoning
8.0%
31.7%
−73.4
SimpleQA VerifiedKnowledge
33.2%
75.6%
−56.1
FrontierMath (Tiers 1–3)Math
42.8%
93.7%
−54.3
WeirdMLCoding
49.9%
92.9%
−46.3
SciCodeCoding
47.3%
56.5%
−14.7
GPQA DiamondReasoning
88.8%
95.8%
−9.8
Mock AIME 2024–2025Math
93.3%
100.0%
−6.7

Bars are distances to the best published result on each test (percent scores above chance, Arena ratings as win rate against the leader, open-ended values on a log scale). Raw values as the publishers report them.

Open GPT-6 Astra

Scores

Every test, every source

The best published run per benchmark, the raw value as the publisher reports it, who produced the number, how close it comes to the best published result on that test, and where the model sits among every model scored on it. Percent scores are read above chance, Arena ratings as a win rate against the board leader, open-ended values on a log scale. A test with too few models to compare against is shown but not counted.

Category means

Reasoning57.5 · 2/6
Knowledgebeside43.9 · 1/1
Coding48.2 · 3/7
Math41.3 · 4/5
Agentic0.0 · 1/8
Multimodalbeside79.6 · 1/1
Human preferencebeside81.7 · 1/1
Long context—

Mean · tests taken of the category's comparable tests. A faded category either rests on fewer than half of them or is a single-test category (marked “beside”), which is shown next to the Index and never averaged into it.

GPQA DiamondEpoch-runReasoning

88.8%

90.2% of the best (95.8%), above chance
56th percentile of 78 · n = 198

high

Source: Epoch AI — AI Benchmarking Hub

CritPtLeaderboardReasoning

8.0%

24.8% of the best (32.3%)
48th percentile of 78

high

Source: Epoch AI — AI Benchmarking Hub

SimpleQA VerifiedEpoch-runKnowledge

33.2%