GPT-6 Astra

Official

openai/gpt-6-astra

GPT-6 Astra is OpenAI's new flagship, built for the hardest reasoning, coding, and long-running agent work. It takes text and image input, holds a 1.05M-token context window (922K max input), writes up to 128K output tokens, and exposes five reasoning-effort levels from low through max. Prompt caching is supported, and cache reads bill at one tenth of input.

1050K 上下文視覺理解推理工具呼叫程式編碼提示詞快取付費快取寫入

定價

來源：openai_2026_09_gpt6_x0.10 · 核驗於 2026-09-05

輸入官方 $10 / 百萬 tokens

$1/ 百萬 tokens1折

輸出官方 $50 / 百萬 tokens

$5/ 百萬 tokens1折

快取讀取 $0.1/M · 快取寫入 $1.25/M

檔位	輸入	快取讀取	快取寫入	輸出
Standard ≤272K	$1	$0.1	$1.25	$5
Standard >272K	$2	$0.2	$2.5	$7.5

成本估算

輸入 tokens輸出 tokens

預估費用$0.015

支援協定

OpenAI Chat CompletionsStatus: 可用: 支援串流https://api.onehop.ai/v1
OpenAI ResponsesStatus: 可用: 支援串流https://api.onehop.ai/v1
Anthropic MessagesStatus: 不支援—
Google Vertex AIStatus: 不支援—
OpenAI ImagesStatus: 不支援—
OpenAI SoraStatus: 不支援—

在對話台試一下

呼叫 ID

API 呼叫時用任一 ID 都可以。

OneHop 名稱openai/gpt-6-astra

可用模型 ID

請求中的 model 可以直接使用這些名稱。

gpt-6openai/gpt-6

動手試一下

把 ONEHOP_KEY 佔位符替換為你的 API Key。建立 →

from openai import OpenAI

client = OpenAI(
    base_url="https://api.onehop.ai/v1",
    api_key="<ONEHOP_KEY>",
)

completion = client.chat.completions.create(
    model="openai/gpt-6-astra",
    messages=[{"role": "user", "content": "What is the meaning of life?"}],
)
print(completion.choices[0].message.content)

base_url: https://api.onehop.ai/v1

檔位

輸入

快取讀取

快取寫入

輸出

Standard ≤272K

$0.1

$1.25

Standard >272K

$0.2

$2.5

$7.5

from openai import OpenAI client = OpenAI( base_url="https://api.onehop.ai/v1", api_key="<ONEHOP_KEY>", ) completion = client.chat.completions.create( model="openai/gpt-6-astra", messages=[{"role": "user", "content": "What is the meaning of life?"}], ) print(completion.choices[0].message.content)