Grok
Grok is the model family from xAI, whose developer docs now brand as SpaceXAI. The models hub we fetched on 2026-08-17 marks Grok 4.6 (grok-4-6) as Latest: flagship for code and everything else, agentic tool calling, configurable reasoning, 500k context, $2.00 / $6.00 per 1M input/output tokens. Sibling APIs on the same page include Voice (from $0.05/min) and Imagine image/video (images from $0.02, video from $0.05/sec). This family page should send readers to the current flagship, not freeze on Grok 4.1 or an unverified 1M-context / AIME 95.8% claim.
For the deep 4.6 writeup see Grok 4.6. For the retired beta id see xAI: Grok Beta.
Key Features
- Current flagship grok-4.6: 500k context, configurable reasoning, agentic tools, listed at $2 / $6 per 1M tokens.
- X search and web tools: Docs still list Web Search, X Search, code execution, and MCP.
- Imagine: Image generation/editing and video generation/editing/extension on the same developer site.
- Voice: Real-time speech, TTS, STT.
- Grok Build / CLI: Separate product surface for coding, next to the API.
Use Cases
- Coding and long agents on the 500k 4.6 window.
- Live X/web questions via official search tools, not by assuming the weights contain the firehose.
- Image/video drafts on Imagine when you already have an xAI key.
Pricing
Official models hub (accessed 2026-08-17):
| Surface | Published rate |
|---|---|
| grok-4.6 | $2.00 / 1M input, $6.00 / 1M output |
| Voice agent | from $0.05 / min |
| TTS | $15.00 / 1M chars |
| Imagine image | from $0.02 / image (1K/2K) |
| Imagine video | from $0.05 / sec (480p/720p/1080p) |
Consumer Grok app pricing is a different SKU (X Premium+ / SuperGrok). Recheck docs.x.ai/developers/models.
Getting Started
- Open the models hub and confirm the Latest id.
- Create an API key and call
grok-4-6. - For coding, also look at Grok Build / CLI in the same docs nav.
- Read Grok 4.6 before you copy older 4.1 numbers.
Frequently Asked Questions
Is Grok still 1M context?
The live 4.6 card we fetched says 500k tokens. Do not keep 1M without a current table.
SpaceXAI or xAI?
Developer docs use SpaceXAI in the chrome and still live at docs.x.ai. Both names appear first-party.
App vs API?
The Grok consumer app on X is not the same billable object as grok-4-6 API tokens.
Alternatives
- Grok 4.6: Full flagship entry.
- GPT-5.6 Sol: OpenAI frontier.
- Claude Opus 5: Anthropic coding flagship.
Tips
- Quote the models hub date. xAI has been retiring older ids (docs mention a May 15 retirement).
- Keep Imagine video rates next to Veo 3 and MiniMax H3, not inside the chat-model paragraph.
- Fix any leftover YouTube embeds that used a full youtu.be URL as a videoId.
Conclusion
Grok is a family. In August 2026 the official Latest card is grok-4.6 at 500k and $2/$6. Start at the SpaceXAI models hub, then use the dedicated 4.6 page for depth.
Comments
No comments yet. Be the first to comment!
Related Tools
Grok 4.6
x.ai
SpaceXAI's frontier 1.5T reasoning model with 500K-token context, agentic coding strength, and real-time X data at roughly half the price of rivals.
xAI: Grok Beta
docs.x.ai/developers/models
Previous xAI Grok Beta snapshot. Rechecked: live docs lead with grok-4.6. Not a 2026 reasoning default.
GPT-5.2
developers.openai.com/api/docs/models/gpt-5.2
OpenAI's previous frontier model: 400k context, 128k max out, $1.75/$14 per 1M tokens. Docs now recommend GPT-5.6 Sol.
Related Insights

Grok Bot and Hermes Bot: one person finally gets a think tank and a secretariat
Grok Bot now ships with Cursor Pro+. Hermes Bot runs on a VPS. They are not smarter chat boxes. The think tank advises, the secretariat executes, and you still make the call.
What Locks Codex Is the Picker, Not the Models
You already pay for OpenCode Go, Grok, and Z.ai, but the Codex picker still shows mostly GPT. The community project codex-router does not teach another install ritual. It puts subscriptions you already bought back into the selector, keeps keys on the machine, and leaves native GPT alone.
Hook OpenCode Go into Codex on Windows. Do Not Open a Second Toolkit.
A ChatGPT-signed Codex desktop app still shows mostly GPT in the picker. On Windows, enable only OpenCode Go and the Grok, GLM, Kimi, DeepSeek, and MiniMax models you already pay for appear in the same selector. Keys stay local. Native GPT stays put.