Grok 4.3 (flagship model)

xAI's reasoning-first flagship released to API on 30 April 2026. A 1M token context window, native video input, and a permanent reasoning state that 'thinks' before responding. Scores 53 on the Artificial Analysis Intelligence Index, below GPT-5.5 (60) and Gemini 3.1 Pro (57) but at a fraction of the price. Accessed via grok.com, X apps, or the API.

• 1M token context window
• Native video input up to five minutes at 1080p, with speaker segmentation and motion reasoning
• Generates downloadable PDFs, populated spreadsheets, and PowerPoint decks directly from chat
• Permanent reasoning state for multi-step tasks
• API pricing at $1.25 per 1M input tokens and $2.50 per 1M output tokens

• Summarise an hour-long site walk video into a defects list with timestamps
• Generate a first-draft pitch deck in PowerPoint from a meeting transcript
• Run a real-time scan of X for what the trade press is saying about a competitor launch
• Build a research brief that mixes public web sources with X commentary on the same topic
• Cheap bulk classification or extraction work where Claude or GPT would be three times the cost

Below Claude 4.6 and GPT-5.5 on serious reasoning benchmarks. No Business Associate Agreement (BAA), no UK enterprise data processing commitments comparable to Anthropic or OpenAI. ADL bias audit placed it last out of six major models. 'Anti-woke' alignment posture produces unpredictable outputs that are a liability for client-facing work. Enterprise adoption sits a long way behind Claude and ChatGPT.

API: $1.25 per 1M input tokens, $2.50 per 1M output tokens. Consumer plans: free tier (limited), X Premium £8/month (basic Grok in X), SuperGrok Lite £10/month, SuperGrok £30/month, X Premium+ £40/month, SuperGrok Heavy £300/month (full 4.3 access and multi-agent reasoning).