---
last_updated: 2026-11-22
source_url: https://opper.ai/xai/grok-4-6
title: "Grok 4.6 by xAI"
description: "Grok 4.6 is xAI's August 2026 frontier release, tuned for long-running agents and coding with a 500K context window and vision input."
---

# Grok 4.6

By [xAI](https://opper.ai/xai).

Released 12 Aug 2026, 5 days after [Grok Imagine Image 2.0](https://opper.ai/xai/grok-imagine-image-2-0) · [Announcement](https://x.ai/news/grok-4-6) · [All xAI releases](https://opper.ai/model-releases/xai)

Grok 4.6 arrived on August 12, 2026 as the successor to Grok 4.5, and xAI held the underlying foundation constant rather than pretraining anew. The improvement came from a longer supplemental training run, regenerated supervised fine-tuning trajectories and reinforcement learning inside agentic environments, and the model now checks and verifies its own work more often before moving on to the next step.

The gains concentrate on agentic work, with DeepSWE moving from 54 to 65.9 and APEX-Agents from 47.1 to 57.5 against Grok 4.5. It holds a 500K token context window with vision input, native tool calling and structured output, and carries a knowledge cutoff of February 1, 2026, which makes it a fit for coding agents and long-horizon automation that have to stay coherent across many steps.

## Available routes

Grok 4.6 runs on 5 different routes through the Opper gateway. Prices are per 1M tokens. Some routes also sell priority processing: the same model at another price, picked per request with service_tier. [How service tiers work](https://docs.opper.ai/build/gateway/service-tiers). Its routes and prices, live from the catalogue, are public JSON with no API key needed at https://api.opper.ai/v3/models?q=xai%2Fgrok-4.6.

| Provider | Model id | Region | Zero data retention | Training | Input | Output | Cache read |
| --- | --- | --- | --- | --- | --- | --- | --- |
| [Azure](https://opper.ai/provider/azure) | `azure:global/grok-4.6` | Multi-region | Abuse monitoring | No | $2.00 | $6.00 | $0.50 |
| [xAI](https://opper.ai/provider/xai) | `xai-zdr/grok-4.6` | Multi-region | Yes | No | $2.00 | $6.00 | $0.50 |
| [xAI](https://opper.ai/provider/xai), Priority | `xai-zdr:priority/grok-4.6` | Multi-region | Yes | No | $4.00 | $12.00 | $1.00 |
| [xAI](https://opper.ai/provider/xai) | `xai-zdr:us/grok-4.6` | US (United States) | Yes | No | $2.20 | $6.60 | $0.55 |
| [xAI](https://opper.ai/provider/xai) | `xai/grok-4.6` | Multi-region | Abuse monitoring | No | $2.00 | $6.00 | $0.50 |
| [xAI](https://opper.ai/provider/xai), Priority | `xai:priority/grok-4.6` | Multi-region | Abuse monitoring | No | $4.00 | $12.00 | $1.00 |
| [xAI](https://opper.ai/provider/xai) | `xai:us/grok-4.6` | US (United States) | Abuse monitoring | No | $2.20 | $6.60 | $0.55 |

## Key info

- Model id: `xai/grok-4.6`
- Input: Text, Image
- Output: Text
- Features: Reasoning, Tools, Structured output, PDF input
- Context window: 500K
- Max output: n/a
- Input price: $2.00 per 1M tokens ($0.50 cache read)
- Output price: $6.00 per 1M tokens
- Released: 12 Aug 2026
- US residency available on xAI
- Zero data retention on pay-as-you-go
- No training by default

## Data handling per route

Each route hosting Grok 4.6 has its own privacy posture, residency, and GDPR terms. Postures are maintained by Opper with a last-verification timestamp.

### Azure — Multi-region

Zero data retention with abuse monitoring: nothing is logged or used for training; content flagged by abuse monitoring is held for review. No training on customer data. GLOBAL; SCCs.

- Zero data retention: With abuse monitoring, flagged content held.
- Training: No training on customer data.
- Logging: None
- Abuse monitoring: On by default, holds flagged content, a person can view samples
- Transfer mechanism: SCCs

Verified by Opper on 2026-09-16.

### xAI — Multi-region

Zero data retention: nothing is logged, held for abuse monitoring or used for training. No training on customer data. GLOBAL; SCCs.

- Zero data retention: Yes. Nothing is logged, held for abuse monitoring or used for training.
- Training: No training on customer data.
- Logging: None
- Abuse monitoring: On by default, keeps nothing
- Caching: Content cached for replay
- Transfer mechanism: SCCs

Verified by Opper on 2026-10-15.

### xAI — United States

Zero data retention: nothing is logged, held for abuse monitoring or used for training. No training on customer data. US; SCCs.

- Zero data retention: Yes. Nothing is logged, held for abuse monitoring or used for training.
- Training: No training on customer data.
- Logging: None
- Abuse monitoring: On by default, keeps nothing
- Caching: Content cached for replay
- Transfer mechanism: SCCs

Verified by Opper on 2026-10-15.

### xAI — Multi-region

Zero data retention with abuse monitoring: prompts and outputs are kept for up to 30 days for abuse monitoring only, and are not used for training. No training on customer data. GLOBAL; SCCs.

- Zero data retention: With abuse monitoring, 30-day window.
- Training: No training on customer data.
- Logging: Abuse monitoring (30-day retention)
- Abuse monitoring: On by default, keeps nothing
- Caching: Content cached for replay
- Transfer mechanism: SCCs

Verified by Opper on 2026-10-15.

### xAI — United States

Zero data retention with abuse monitoring: prompts and outputs are kept for up to 30 days for abuse monitoring only, and are not used for training. No training on customer data. US; SCCs.

- Zero data retention: With abuse monitoring, 30-day window.
- Training: No training on customer data.
- Logging: Abuse monitoring (30-day retention)
- Abuse monitoring: On by default, keeps nothing
- Caching: Content cached for replay
- Transfer mechanism: SCCs

Verified by Opper on 2026-10-15.

## Best for

Where Grok 4.6 ranks among the models on Opper with a score for each use case (0 to 100).

- [Finance](https://opper.ai/models?use=finance): 47.6, #19 of 53 (Finance and accounting)
- [Healthcare](https://opper.ai/models?use=healthcare): 44.1, #19 of 50 (Healthcare and medical)
- [Legal](https://opper.ai/models?use=legal): 43.5, #20 of 53 (Legal tasks and reasoning)
- [Agents](https://opper.ai/models?use=agents): 44.0, #31 of 93 (Terminal-Bench 4.0 and AutomationBench)
- [Coding](https://opper.ai/models?use=coding): 18.8 (FrontierCode 2.0 and Terminal-Bench 4.0)

## Benchmarks

Independent benchmark scores via [Artificial Analysis](https://artificialanalysis.ai) ([view on AA](https://artificialanalysis.ai/models/grok-4-6)). Indices and domain scores are out of 100, Omniscience runs from −100 to 100, Briefcase is an Elo rating, and the other evaluation scores are the share of tasks solved.

- Global rank: #29 of 520 LLMs on the [LLM leaderboard](https://opper.ai/llm-leaderboard)
- Tier: Strong
- Output speed: 71 tok/s
- Intelligence Index: 31.0

### Coding

- FrontierCode 2.0: 20%
- Terminal-Bench 4.0: 21%
- SciCode: 56%

### Agentic & knowledge work

- AutomationBench: 67%
- GDPval: 58%
- GDP.pdf: 18%
- Briefcase (Elo): 1539

### Science & reasoning

- Terminal-Bench Science: 6%
- CritPt (hard): 40%
- HLE Diamond: 22%
- Humanity's Last Exam: 44%
- GPQA Diamond: 95%
- Long-context reasoning: 81%

### Knowledge

- Omniscience: 24.9

### By domain

- Finance & accounting: 47.6
- Legal: 43.5
- Healthcare & medical: 44.1
- Engineering: 37.4
- Economics: 42.0
- Strategy & operations: 52.8

## Get started

Call Grok 4.6 through the Opper gateway with one API key. Opper is drop-in compatible with the OpenAI, Anthropic and Google AI SDKs.

Grok 4.6 is a premium model on Opper. Sign up needs no credit card: you get an API key straight away and the free models work in the playground and the API. Add a card to use premium models, pay-as-you-go with no minimum.

### Set it up with your agent

Paste this into a coding agent like Claude Code, Cursor or Codex and it will wire up Opper for you:

```text
Use curl to download, read and follow: https://skills.opper.ai Call Grok 4.6
```

### Or call it directly

OpenAI:

```javascript
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.OPPER_API_KEY,
  baseURL: "https://api.opper.ai/v3/compat",
});

const completion = await client.chat.completions.create({
  model: "xai/grok-4.6",
  messages: [{ role: "user", content: "Hello" }],
});
console.log(completion.choices[0].message.content);
```

Anthropic:

```javascript
import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  apiKey: process.env.OPPER_API_KEY,
  baseURL: "https://api.opper.ai/v3/compat",
});

const message = await client.messages.create({
  model: "xai/grok-4.6",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Hello" }],
});
console.log(message.content);
```

Gemini:

```javascript
import { GoogleGenAI } from "@google/genai";

const ai = new GoogleGenAI({
  apiKey: process.env.OPPER_API_KEY,
  httpOptions: { baseUrl: "https://api.opper.ai/v3/compat" },
});

const response = await ai.models.generateContent({
  model: "xai/grok-4.6",
  contents: "Hello",
});
console.log(response.text);
```

cURL:

```bash
curl https://api.opper.ai/v3/compat/chat/completions \
  -H "Authorization: Bearer $OPPER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "xai/grok-4.6",
    "messages": [{ "role": "user", "content": "Hello" }]
  }'
```

## Use Grok 4.6 in your apps

Pick your tool to see how to run it on Grok 4.6 through Opper.

- [Claude Code](https://opper.ai/apps/claude-code/grok-4-6)
- [Claude Cowork](https://opper.ai/apps/claude-cowork/grok-4-6)
- [Codex](https://opper.ai/apps/codex/grok-4-6)
- [DeepSeek Harness](https://opper.ai/apps/deepseek-harness/grok-4-6)
- [GitHub Copilot CLI](https://opper.ai/apps/github-copilot-cli/grok-4-6)
- [Aider](https://opper.ai/apps/aider/grok-4-6)
- [Goose](https://opper.ai/apps/goose/grok-4-6)
- [Droid](https://opper.ai/apps/droid/grok-4-6)
- [OpenCode](https://opper.ai/apps/opencode/grok-4-6)
- [Pi](https://opper.ai/apps/pi/grok-4-6)
- [OpenClaw](https://opper.ai/apps/openclaw/grok-4-6)
- [Hermes](https://opper.ai/apps/hermes/grok-4-6)
- [Copilot for Obsidian](https://opper.ai/apps/obsidian-copilot/grok-4-6)
- [OmniRoute](https://opper.ai/apps/omniroute/grok-4-6)
- [Cursor](https://opper.ai/apps/cursor/grok-4-6)
- [Cline](https://opper.ai/apps/cline/grok-4-6)
- [Continue](https://opper.ai/apps/continue/grok-4-6)
- [Kilo Code](https://opper.ai/apps/kilo-code/grok-4-6)
- [GitHub Copilot](https://opper.ai/apps/github-copilot/grok-4-6)
- [Zed](https://opper.ai/apps/zed/grok-4-6)
- [JetBrains AI Assistant](https://opper.ai/apps/jetbrains-ai-assistant/grok-4-6)
- [Voidleap Code](https://opper.ai/apps/voidleap-code/grok-4-6)
- [n8n](https://opper.ai/apps/n8n/grok-4-6)
- [LibreChat](https://opper.ai/apps/librechat/grok-4-6)
- [Open WebUI](https://opper.ai/apps/open-webui/grok-4-6)
- [See all apps](https://opper.ai/apps)

## Other models from xAI

- [Grok 4.7](https://opper.ai/xai/grok-4-7)
- [Grok 4.5](https://opper.ai/xai/grok-4-5)
- [Grok 4.3](https://opper.ai/xai/grok-4-3)
- [Grok 4.20 Non-Reasoning](https://opper.ai/xai/grok-4-20-0309-non-reasoning)
- [Grok 4.20 Non-Reasoning](https://opper.ai/xai/grok-4-20-non-reasoning)
- [Grok Voice](https://opper.ai/xai/grok-voice-latest)
