Advertise here
#33

Grok 4.3

X AI Release: 2026-05-01 Tested on: 2026-05-22 00:32 x-ai/grok-4.3::medium
~1.5T total (~150B active)MoEClosedEstimated

Summary

Grok 4.3 scores 7.8 on AI BENCHY and ranks #33. It has 10.0 reliability, a 75.0% pass rate, $0.593 total cost, and 49.23s average response time.

Model facts

Researched on 2026-08-12

Estimated
Parameters
~1.5T total (~150B active)
Architecture
MoE
Availability
Closed
License
-

Best estimate from public evidence; the vendor did not disclose every value. xAI does not disclose the architecture or count.

Parameter source Availability source

Score

7.8

Consistency

8.4

Reliability

10.0

Total Cost (Current Price)

$0.593

Total Output Tokens

216,413

Total Input Tokens

0

Input Price

$1.250 / 1M

Output Price

$2.500 / 1M

Tests Correct

Wrong Tests: 7

Attempt pass rate: 75.0%

Flaky tests

4

Flaky tests had mixed outcomes across runs (at least one pass and one fail).

Response Time (avg)

49.23s

Response Time (max): 216.69s

Response Time (total): 984.54s

Wrong answer: 4 Did not follow instructions: 2 Extra formatting: 1

Hamster playing table tennis

Prompt: Create a detailed SVG illustration of a hamster playing table tennis.

#33 xAI: Grok 4.3

medium
Cost
$0.009
Time
19.0s
Tokens
3,661 tok

Run history

Tested on Score Reliability Tests Correct Total Cost Compare
2026-08-14 02:54 Re-test 7.0 10.0 $0.741 Compare
2026-07-16 23:07 New test added 7.1 10.0 $0.779 Compare
2026-06-04 14:11 New test added 7.6 10.0 $0.614 Compare
2026-05-22 00:32 Re-test 7.8 10.0 $0.593 Current run
2026-05-01 00:40 Initial run 8.2 10.0 $0.517 Compare

Run comparison

RunBenchmark coverageScoreConsistencyReliabilityTests CorrectFlaky testsTotal Output TokensTotal Input TokensTotal CostResponse Time (avg)
2026-05-22 00:32 · Re-test60/60 attempts7.88.410.013/204216,4130$0.59349.23s
2026-08-14 02:54 · Re-test66/66 attempts7.08.210.012/225225,904140,244$0.74144.21s
Difference—+0.7+0.20.0+1-1-9491-140244-$0.148+5016ms

Benchmark coverage differs: 60/60 attempts (Target: 3 repeats per test) versus 66/66 attempts (Target: 3 repeats per test). Totals and repeat-sensitive metrics are not directly comparable.

These two runs used different benchmark suites, so the deltas reflect both model changes and suite changes.

Charts

Choose the first model, then click a second model to open a side-by-side page.

Total Output Tokens

Score vs Total Output Tokens

Quick Compare

Grok 4.3mediumvsMercury 2.5highGrok 4.3mediumvsGPT-5.6 SolnoneGrok 4.3mediumvsSolar Pro 4highGrok 4.3mediumvsDeepSeek V3.2mediumGrok 4.3mediumvsStep 3.7 FlashlowGrok 4.3mediumvsGLM 5.1mediumGrok 4.3mediumvsMuse Glimmer 30BlowGrok 4.3mediumvsGPT-5.5noneGrok 4.3mediumvsQwen3.5-122B-A10BmediumGrok 4.3mediumvsMercury 2medium

Category Breakdown

Category Score Consistency Tests Correct
Anti-AI Tricks 10.0 10.0
Coding 7.4 6.5
Combined 10.0 10.0
Data parsing and extraction 10.0 10.0
Domain specific 5.3 7.2
General Intelligence 5.4 2.5
Instructions following 9.8 10.0
Puzzle Solving 5.9 7.2
Tool Calling 10.0 10.0
Trivia 3.0 10.0

Compared models