Transparent pricing with no hidden fees. Pay as you go.
grok-4.6grok-4.6| Model | Spec | PoYo price | Official price | You save |
|---|---|---|---|---|
| grok-4.6 | Input | $1.60/1M tokens 320 credits /1M tokens | $2.00/1M tokens Official | 20% |
Output (including reasoning tokens) | $4.80/1M tokens 960 credits /1M tokens | $6.00/1M tokens Official | 20% |
* Actual fees are based on the final output.
Complete guide to using Grok 4.6 API for Long-Running Agents
Grok 4.6 is xAI's frontier model for coding, engineering, knowledge work, and ambitious interactive or visual projects. It can stay with complex tasks across many steps, research unfamiliar domains, work through a codebase, and self-check results before moving forward.
Use grok-4.6 through /v1/chat/completions or /v1/responses. PoYo charges 320 input and 960 output credits per million tokens—$1.60 and $4.80, exactly 20% below xAI's $2 and $6 standard short-context rates.
The official xAI launch focuses on sustained agent work and stronger first passes for interactive and visual projects.
01
Grok 4.6 was trained for knowledge work, general coding, kernel optimization, web development, computer-aided design, and other agentic environments. xAI reports more self-testing and verification on longer trajectories.
02
xAI designed Grok 4.6 to take on more ambitious interactive and visual work. It can carry a broad brief through planning and implementation, then produce a more substantial first version of an application or polished work artifact.
03
Connect Grok 4.6 through either PoYo endpoint and keep the same public model ID from quick chat tests to long-running agent workflows. Combine streaming, tool calls, reasoning control, and context compaction without maintaining a second integration.
Sustain research, planning, implementation, testing, and refinement across many steps without losing the objective.
Navigate repositories, diagnose defects, implement changes, run tools, and verify results in agentic coding workflows.
Turn broad product ideas into substantial first versions with coherent structure, interactions, and visual language.
Analyze reports and complex information, synthesize evidence, and create polished work artifacts within a 500K context.
Combine function calling with web search, X search, code execution, structured outputs, and external application tools.
Choose low, medium, high, or xhigh reasoning effort to match latency and depth to each task.
xAI's official high-reasoning evaluations compare Grok 4.6 with Grok 4.5, GPT-5.6 Sol Max, and Fable 5 Max across composite intelligence, coding agents, terminal work, and professional knowledge tasks.
| Benchmark | Grok 4.6 High | Grok 4.5 High | GPT-5.6 Sol Max | Fable 5 Max |
|---|---|---|---|---|
| AA Intelligence Index | 61 | 56 | 61 | 62 |
| GDPVal-AA v2 | 1753 | 1526 | 1728 | 1741 |
| CursorBench v3.2 | 69.9% | 66.7% | 67.2% | 70.5% |
| DeepSWE v1.1 | 65.9% | 54% | 73% | 70% |
| FrontierCode v1.1 Extended | 61.3% | 56.6% | 60.6% | 63.6% |
| APEX-Agents | 57.5% | 47.1% | 56.7% | 59.2% |
| Terminal-Bench v3.0 | 26% | 15.7% | 34.6% | 34.1% |
| APEX-SWE | 56.4% | 53.6% | — | 58.8% |
| AA-Briefcase | 1577 | 1313 | 1502 | 1574 |
| Harvey LAB (Vals) | 15.8% | 12.9% | 2.5% | 11.3% |
Official xAI launch evaluations, published August 12, 2026. Third-party model scores are the best self-reported or publicly available results used by xAI.
1. Create an API key. Sign in and create a PoYo key.
2. Choose an endpoint. Use POST /v1/chat/completions with a messages array, or POST /v1/responses with input. Both support streaming.
3. Set the model. Send grok-4.6 as the model ID. For reliable prompt-cache affinity, keep a stable conversation identifier when your client supports it. Read the PoYo API documentation.
grok-4.6 model ID with either supported endpoint.POST /v1/chat/completions and POST /v1/responses, including streaming and non-streaming requests.messages history. Responses uses input, can continue with prior response state, and supports context compaction for long-running workflows.Use Chat Completions or Responses with the same public model ID and PoYo credential.
See input and output rates in both credits and US-dollar equivalents against official short-context pricing.
Evaluate Grok alongside other model providers without adding another gateway integration.
Switch endpoints, test streaming, review service status, and inspect token usage before production rollout.