xAI (Grok) / grok-4-20-non-reasoning
Grok 4 20 Non-Reasoning stands out for Grok variant prioritizing low latency over long reasoning and Strict instruction following for frequent tool calls. It is best suited to realtime agent interaction and responsive tool workflows.
Context
128K
token
Input
$0.20
/ 1M tokens
Output
$0.50
/ 1M tokens
Quality
#42
Elo 1240
Overview
Grok 4 20 Non-Reasoning stands out for Grok variant prioritizing low latency over long reasoning and Strict instruction following for frequent tool calls. It is best suited to realtime agent interaction and responsive tool workflows.
Tool calling
Can call tools or functions, so product flows can connect model output to backend actions.
Streaming
Supports streamed responses for chat, assistants, and interfaces that need visible incremental output.
Model maker
xAI (Grok)
Called through SeaLink's unified account, balance, and OpenAI-compatible API.
Protocol
OpenAI-compatible Chat Completions
Base URL: https://api.sealink.io/v1
Pricing
Light production run
100K input + 25K output
$0.0325
Higher-volume run
1M input + 250K output
$0.325
The final charge follows generated usage and the applicable selector, such as resolution, duration, input type, or provider unit.
API
model
grok-4-20-non-reasoningbase_url
https://api.sealink.io/v1Auth
Bearer $SEALINK_API_KEYcurl https://api.sealink.io/v1/chat/completions \-H "Authorization: Bearer $SEALINK_API_KEY" \-H "Content-Type: application/json" \-d '{"model": "grok-4-20-non-reasoning","messages": [{"role": "user", "content": "Hello."}]}'
from openai import OpenAIclient = OpenAI(base_url="https://api.sealink.io/v1", api_key="<your-sealink-key>")resp = client.chat.completions.create(model="grok-4-20-non-reasoning",messages=[{"role": "user", "content": "Hello."}],)print(resp.choices[0].message.content)
import OpenAI from "openai";const client = new OpenAI({baseURL: "https://api.sealink.io/v1",apiKey: process.env.SEALINK_API_KEY,});const resp = await client.chat.completions.create({model: "grok-4-20-non-reasoning",messages: [{ role: "user", content: "Hello." }],});console.log(resp.choices[0].message.content);