Skip to main content
OpenAI

GPT-6.2 Astra

GPT-6.2 Astra is a chat model from OpenAI, tracked on this page with third-party throughput and latency measurements alongside provider pricing.

Input from
$8.00 / 1M tokens
on OpenRouter · 3 providers
Output from
$40.00 / 1M tokens
on OpenRouter

API Pricing

Cheapest input on OpenRouter · cheapest output on OpenRouter
ProviderInput / 1MOutput / 1MCached / 1M
$4.00$20.00$0.200
$8.00$40.00$0.400
$8.00$40.00$0.400
$8.00$40.00$0.400

Prices updated daily. Last check: Nov 18, 2026

Compare API pricing for every OpenAI model →

GPT-6.2 Astra pricing by provider

Input, output, and batch rates, plus alternatives, for one provider at a time.

Performance & Benchmarks

Source: Artificial Analysis →
Intelligence
62.1 / 100
Output Speed
77.7 t/s
Latency (TTFT)
435.3s

Benchmarks measured Nov 2026. Scores are independent evaluations, not vendor-reported.

Model Details

General

Creator
OpenAI
Modalities
Text

Capabilities

Strengths & Limitations

Strengths

  • Chat-formatted interface, suitable for drop-in use in message-based applications
  • Measured output throughput of approximately 42.1 tokens per second (Artificial Analysis)
  • Independent third-party benchmark data available rather than vendor-reported figures only
  • Part of OpenAI's GPT line, which is widely supported by existing SDKs, frameworks, and tooling
  • Pricing from multiple providers is tracked side by side in the table on this page
  • Streaming-capable response pattern makes the throughput figure directly relevant to perceived speed

Limitations

  • Time to first token of roughly 3,540 ms makes it a poor fit for latency-sensitive, interactive turn-taking
  • Output throughput of about 42 tokens per second is moderate, so long responses take noticeable wall-clock time
  • Context window size is not confirmed in our database, so maximum prompt length should be verified with the provider
  • Input and output modalities beyond text chat are not tracked for this entry
  • We do not currently track task benchmark scores (reasoning, coding, or multilingual) for this model

Key Features

•Chat/completions-style message interface
•Measured output speed of ~42.1 tokens/second
•Measured time to first token of ~3,540 ms
•Benchmark data sourced from Artificial Analysis
•OpenAI GPT-family model
•Multi-provider pricing comparison on this page

About GPT-6.2 Astra

GPT-6.2 Astra is a chat-oriented large language model created by OpenAI and listed in the computeprices.com catalog as part of the GPT series. Our database records it as a conversational (chat) model, meaning it is designed to take message-formatted prompts and return text responses, as opposed to an embedding, reranking, or media-generation model. The measurements we track for GPT-6.2 Astra come from Artificial Analysis: an output throughput of roughly 42.1 tokens per second and a time to first token of about 3,540 milliseconds. That combination — a multi-second delay before the first token, followed by a moderate streaming rate — is characteristic of models that do additional work before emitting output, and it is the main performance signal we can report with confidence for this entry. Other attributes such as the context window, supported input modalities, and tool-calling behavior are not currently confirmed in our database, so we do not assert them either way. In practice, the way to evaluate GPT-6.2 Astra against alternatives is to weigh those latency and throughput figures against the per-provider pricing shown in the table on this page, and to confirm current specifications directly with the serving provider or OpenAI's documentation before committing to an integration.

Common Use Cases

GPT-6.2 Astra is a chat model, so it fits text-in, text-out workloads such as assistants, document question answering, drafting and rewriting, summarization, and backend content generation. Given the recorded time to first token of roughly 3.5 seconds, it is better suited to asynchronous or batch-style jobs — pipelines, scheduled content generation, offline analysis, and background agents — than to voice interfaces, autocomplete, or other uses where the user is waiting on the first character. The ~42 tokens/second output rate is workable for short and medium-length replies; for very long generations, budget for the additional wall-clock time or stream results progressively into the UI. Because the context window and modality support are not confirmed in our data, check the provider's documentation before designing around large prompts or non-text inputs.

Frequently Asked Questions

How much does GPT-6.2 Astra cost?

Pricing depends on the provider serving the model and on the pricing type — input tokens, output tokens, and any cached or batch rates are billed separately and change over time. See the pricing table on this page for the current figures we track across providers.

What is GPT-6.2 Astra best used for?

It is a chat model, so it suits assistant-style conversations, summarization, drafting, and text generation pipelines. Its measured ~3.5 second time to first token makes it a better match for asynchronous and batch workloads than for real-time interactive experiences.

How fast is GPT-6.2 Astra?

Artificial Analysis measures it at approximately 42.1 output tokens per second with a time to first token of about 3,540 milliseconds. The first-token delay dominates short requests, while the token rate dominates long generations.

What is the context window of GPT-6.2 Astra?

We do not currently have a confirmed context window for this model in our database. Check the documentation of the provider you plan to use before building around long prompts.

Who created GPT-6.2 Astra?

OpenAI. It is listed in our catalog as part of the GPT family of chat models.