X Introduces Grok-4 Flagship Model with 256k Context Window and New Enterprise Features
What changed
grok-2-vision-1212 multimodal model discontinued
grok-4 flagship model introduced with 256k context window at $3.00 input/$15.00 output per million tokens
grok-4-fast-non-reasoning model introduced with 2M context window at $0.20 input/$0.50 output per million tokens
grok-code-fast-1 model introduced with 256k context window at $0.20 input/$1.50 output per million tokens
grok-3-mini context window increased from 131072 to 132k tokens
Enterprise section introduced with SSO, audit logging, authorization controls, compliance features (SOC 2 Type 2, GDPR, CCPA, Zero Data Retention), and data residency
Search feature changed from 'Coming soon' to 'New' with real-time capability now available
grok-3-beta renamed to grok-3 (dropped beta designation)
grok-4-fast-reasoning model introduced with 2M context window at $0.20 input/$0.50 output per million tokens
grok-3 context window increased from 131072 to 132k tokens
grok-3-mini-beta renamed to grok-3-mini (dropped beta designation)
Before and after
Jul 6, 2025 vs Oct 11, 2025grok-2-vision-1212 multimodal model discontinued
Sign in to view the full comparison
See the before/after screenshots and every marked-up change.
More from X
Full history →- Week 35, 2026·improvedX drops grok-4.5, adds Imagine image/video models and free Playground perk
- Week 32, 2026·packagingx.ai: Speech-to-Speech Voice API pricing changed from flat rate to starting-at
- Week 29, 2026·packagingX swaps flagship model — grok-4.5 costs more, sees less, drops coding model
- Q2 2026·improvedxAI: Grok lineup replaced (grok-4.3, grok-build-0.1), smaller context windows
- Week 22, 2026·packagingX launches Grok coding model and image/video/voice API pricing