The Open-Source AI Stack
RSS
All models

Models · grok

Grok 4.3

Proprietary xAI · — · Proprietary

xAI's most-intelligent-and-fastest model per the lab, released April 30 2026 after a Grok 4.3 Beta drop on April 17 for SuperGrok Heavy users. Native multimodal video understanding and direct PDF, spreadsheet, and PowerPoint generation from chat. Reasoning is now always-on rather than toggleable.

Cost

$1.25 / Mtok input
$2.50 / Mtok output

xAI API · as of 2026-05-21

source ↗

Speed

88.1 tok/sec output
10967 ms TTFT

· as of 2026-05-21

source ↗

Architecture

tokens in Embedding vocab not disclosed × N layers Architecture not disclosed (proprietary or undocumented) Output projection tokens out
Schema-generated from data/models.yaml. Every label is auditable against the model's sources.

Specs

Architecture
unknown
Total params
not disclosed
Active params
not disclosed
Context window
not verified
Attention
unknown
Position encoding
unknown
Post-training
rlhf
OSI-approved
no
Data released
no
Training code
not released

Benchmarks

Each score carries the date it was published; we never infer or interpolate missing scores.

General reasoning

GPQA-Diamond 90.1 as of 2026-05-21 source ↗

Available quantizations

None. The weights are not distributed, so there are no public quantizations.

Notable innovations

  • · Always-on reasoning
  • · Native video understanding
  • · Direct PDF/XLSX/PPTX generation from chat

Lineage

Successor to Grok 4.2; always-on reasoning is the architectural shift.

Derived from

Grok 4.1 2025-11-17

Sources