Model

Grok 4.3 (High)

CLOSED

SpaceXAI · Grok 4.3 family · released Apr 30, 2026

High reasoning-effort tier of Grok 4.3, same underlying weights as the default/low-effort configuration, opted into via the API's reasoning_effort parameter. This is the configuration Artificial Analysis benchmarks by default for Grok 4.3.

ReasoningCodingVisionFunction callingTool useAgentic

2298.5

Elo · rank #35

Parameters

Undisclosed

Active params

Undisclosed

Context

1M tokens

Architecture

Mixture-of-Experts (size undisclosed); high reasoning-effort configuration of the same served weights as Grok 4.3

License

Proprietary

Languages

—

API price (in/out)

$1.25 / $2.5

Modalities

text · vision · video

Benchmark results

Bar shows position within the tracked field; marker = field best

Agents' Last ExamAgents15.5%#19

best: GPT-5.6 · 52.7%

GDPval-AAAgents1500#15

best: Claude Fable 5 · 1932

GPQA DiamondReasoning90.1%#26

best: GPT-5.6 · 94.6%

Humanity's Last ExamReasoning35.0%#33

best: Claude Sonnet 5 · 57.4%

IFBenchReasoning81.3%#4

best: MiniMax M3 · 83.0%

LiveCodeBenchCoding84.5%#26

best: DeepSeek-V4-Pro (Think Max) · 93.5%

MMMUVision83.1%#13

best: Claude Fable 5 · 89.3%

SWE-bench VerifiedCoding71.4%#54

best: Claude Fable 5 · 95.0%

τ²-Bench TelecomAgents97.7%#11

best: Claude Opus 4.6 · 99.3%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$1.25

Output / M tok

$2.5

Grok 4.3 family

Elo progression across releases

Grok 4.3Apr 20261973.3 Grok 4.3 (High)Apr 20262298.5

Resources

DOCShttps://docs.x.ai APIAvailable via SpaceXAI API

API price $1.25/$2.5 · each benchmark row carries its own source badge (see methodology)