ModelPriceWatch.com
Last scan 2026-09-04 Models tracked 244 Providers 31 Cheapest paid Granite 4.0 H Micro $0.017/Mtok in Every price links to its source

GPT-6 Astra

by OpenAI

Current new · 1d flagship expensive tier
Email me when GPT-6 Astra gets cheaper
Availability: Rolling out first to enterprises through OpenAI's Trusted Access Program, with Plus, Pro, Business and Enterprise plan access stated as coming in the following days. This rate is published on OpenAI's public rate card all the same, and the model is served over v1/responses, v1/chat/completions and v1/batch.

Today's price · per 1M tokens

Input

$10.00

per 1M tokens

Output

$50.00

per 1M tokens

Blended

$20.00

blended $/1M — 3:1 weighted input:output

Cached input

$1.00

10% of input — prompt caching

How this price is scoped: OpenAI's rate card prices this model by context length: the tracked figure is the short-context tier, and prompts over 272K input tokens bill at the long-context tier of $20 input / $2 cached input / $25 cache write / $75 output per 1M tokens — 2x input and cache rates and 1.5x output, applied to the full request. Cache writes bill at 1.25x the uncached input rate. Batch and Flex are 50% of Standard; Fast mode is 2x the applicable rate, and Fast is unavailable with EU data residency.

Source: official OpenAI pricing · read MODELPRICEWATCH.COM · 2026-09-04

Price receipt

We read OpenAI's own pricing page on and found gpt-6-astra listed with a price on the same row — that read is where the number above comes from, recorded as a new model. We keep the page text we read; its content hash is 3f8b1fb72f.

Overview

OpenAI's most capable model, released 2026-09-03 and positioned above the GPT-5.6 Sol/Terra/Luna tiers for the hardest end-to-end work: complex reasoning, coding, computer use, research and document creation. 1.05M context window, 922K max input, 128K max output, text + image input, text output, April 2026 knowledge cutoff. $10/$50 per 1M tokens (cached input $1.00). Supports reasoning effort levels low through max, but not the none level, and does not accept custom temperature, top_p or logprobs. Tool calling requires the Responses API.

Capabilities

struck through = not supported
Input 2/5
Text ✓ Image ✓ Audio Video PDF
Output 1/5
Text ✓ Image Audio Video Embedding
Features 4/9
Prompt caching ✓ Reasoning ✓ Coding Fast inference Long context ✓ Open weights Multimodal ✓ Web search Realtime

Benchmark performance

accuracy % · higher is better
Avg benchmark score96%
Perf per $/Mtok4.8
GPQA Diamond
96%
AutoBench *
41.4%
Independent composite scores
  • AA Intelligence Index: 61.2 (Artificial Analysis composite across reasoning, knowledge and coding evals)
  • AA Agentic Index: 51.5 (Tool use, planning, autonomy)
  • AA-Omniscience Index: 43.4 (−100–100)
  • GDPval-AA v2: 1629.3 ELO (Real-world work tasks, human baseline = 1000)
How it stacks up
  • Ranks #1 of 76 benchmarked models by average score
  • Ranks #3 of 176 comparably-measured models by percentile score, across 5 independent measurements
  • Ranks #162 of 176 comparably-tested models by normalized performance per dollar
  • Strongest at GPQA Diamond — 96%, #2 of 64

Source: Vellum LLM Leaderboard · updated · See full rankings →

Specifications

Provider
OpenAI
Context window
1M tokens
Modality
text, image
Parameters
Proprietary
Open source
No — proprietary
Released
Status
Current
Last updated
Tags
reasoningmultimodalflagshipcaching

Availability verified: — listed on OpenAI's own page

Try on OpenAI Official pricing page