> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tess.im/llms.txt
> Use this file to discover all available pages before exploring further.

# Grok 4.6

Grok 4.6 (xAI) is available in Tess for **long-running agents, coding, knowledge work, and polished visual/interactive deliverables**. It is a post-training upgrade on the Grok 4.5 base — same price, stronger agentic coding.

| **Model IDs**<br /><br />`grok-4.6-low` · `grok-4.6-medium` · `grok-4.6-high`                | **Context**<br /><br />500K      | **Provider**<br /><br />xAI     | **Released**<br /><br />12 Aug 2026  |
| :------------------------------------------------------------------------------------------- | :------------------------------- | :------------------------------ | :----------------------------------- |
| **Capabilities**<br /><br /><Icon icon="brain" /><Icon icon="image" /><Icon icon="wrench" /> | **Speed**<br /><br />Medium–High | **Cost**<br /><br />Same as 4.5 | **Intelligence**<br /><br />Frontier |

## What changed vs Grok 4.5

|                           | Grok 4.5                   | Grok 4.6                           |
| ------------------------- | -------------------------- | ---------------------------------- |
| Base                      | Grok 4.5                   | Same base (extended post-training) |
| Context                   | 500K                       | 500K                               |
| DeepSWE v1.1              | 54%                        | **65.9%**                          |
| Terminal-Bench v3.0       | 15.7%                      | **26%**                            |
| APEX-Agents               | 47.1%                      | **57.5%**                          |
| Reasoning in Tess         | low / medium / high        | low / medium / high                |
| Tess credits / 100 tokens | 0.096 in / 0.288 out       | **0.096 in / 0.288 out**           |
| Modalities                | Text + image in → text out | Text + image in → text out         |

Native **reasoning**, **tools** (function calling, web search, code execution), and **image input**. Knowledge cutoff: February 2026.

## Pricing (Tess credits)

Values follow [Models and Costs](/en/models-and-cost) (credits per 100 tokens). Same rate across reasoning levels:

| Model           | Input / 100 tokens | Output / 100 tokens |
| --------------- | ------------------ | ------------------- |
| Grok 4.6 Low    | 0.096              | 0.288               |
| Grok 4.6 Medium | 0.096              | 0.288               |
| Grok 4.6 High   | 0.096              | 0.288               |

> **Screenshot placeholder — model picker:** Capture the chat model selector with **Grok 4.6** (Low / Medium / High) visible.

## Ideal use cases in Tess

1. Long-running agents — multi-step research and execution without losing the thread
2. Agentic coding — codebase changes with +12 pts DeepSWE vs 4.5
3. Knowledge work on business documents
4. Interactive prototypes and polished deliverables
5. Multi-agent setups — High reasoning for the hardest subagents
6. Long sessions (500K context) without switching models

<Tip>
  **Best practices**

  * Start on **Medium**; move to **High** only when the task needs longer tool loops or harder coding.
  * Price matches Grok 4.5 — upgrade the default on existing agents when you want the coding/agent gains at no extra credit cost.
  * Very long prompts on the provider side can cost more; keep context scoped when you can.
</Tip>

See also: [Models and Costs](/en/models-and-cost) · [Introducing Grok 4.6 (xAI)](https://x.ai/news/grok-4-6).