Sulat.com
AI models
Requesty logo

Model details

GPT-6 Astra

GPT-6 Astra is OpenAI's latest flagship language model, positioned as a successor in the company's frontier lineup and branded with the "Astra" codename. The model is rolling out in phases rather than as an immediate general release, with companies participating in OpenAI's application-based cybersecurity program getting first access before broader availability through ChatGPT plans and the OpenAI API. This staged approach reflects the heightened attention OpenAI placed on Astra's capabilities at launch, which included explicit warnings about its advanced cyber potential. The launch framing positions Astra as a notably capable generation rather than an incremental upgrade, consistent with how OpenAI has historically framed major new model tiers.

The early public attention on GPT-6 Astra centers on its strength in cybersecurity-adjacent reasoning, with OpenAI itself flagging advanced cyber capabilities as a defining characteristic of this generation. That emphasis suggests practical fit for organizations working on defensive security research, threat analysis, and automation tasks that require careful multi-step reasoning over complex inputs. Because the initial rollout is gated through a curated cybersecurity cohort, prospective users outside that program will likely encounter Astra later through standard consumer and enterprise channels, giving OpenAI room to observe real-world usage patterns before wider exposure.

Requestygpt-6-astragpt-astra

Quick Info

Powered by
Provider
Requesty
Model key
gpt-6-astra
Release date
Sep 4, 2026
Last updated
Sep 4, 2026
Knowledge cutoff
2026-04-30
Input modalities
Output modalities
Capabilities

Cost

Input token cost
$10.00
Output token cost
$50.00

Limits

Input tokens
922,000 tokens
Output tokens
128,000 tokens
Context window
1,050,000 tokens

Transparent token rates

Compare GPT-6 Astra pricing

Rates are shown per one million tokens. Combined means one million input plus one million output tokens.

Browse this family

Latest news about GPT-6 Astra

Requesty

Coverage

Shattered.io's September 8, 2026 launch summary reports that GPT-6 Astra began rolling out on September 3, 2026 and was widened through the week so that most developers and enterprise subscribers were actively seeking access. OpenAI's safety overview calls Astra "the most capable model we have ever broadly deployed," a The API model identifier developers use is confirmed as `gpt-6-astra`, consistent with OpenAI's own documentation. OpenAI's launch announcement states that GPT-6 Astra is rolling out first to a limited set of organizations and then to all ChatGPT Plus, Pro, Business, and Enterprise users, plus the OpenAI API, Microsoft

Requesty

CoverageBenchmark

MindStudio's independent analysis finds that GPT-6 Astra is "not consistently" better than Claude Fable 5.1, beating it on computer use, terminal workflows, and long-context retrieval while trailing on broader indices. On Artificial Analysis's Intelligence Index, Fable 5.1 scores 66 versus Astra's 61, and on the Coding The critique surfaces a methodology issue: Astra's headline ARC-AGI-3 score of 99.9% uses OpenAI's provider-adapter harness, which preserves reasoning state between actions, whereas a standard neutral harness puts Astra at 62.7%. Astra costs about 2.5x more per token than GPT-5.6 Sol ($10/$50 vs $4/$20), and even with

Requesty

CoverageBenchmark

The llm-stats.com tracker ranks GPT-6 Astra at the top of its composite leaderboard, placing #1 in Coding (1 of 258), Reasoning (1 of 350), Tool Calling (1 of 184), and Vision (1 of 204), with a blended price of approximately $11.9 per million tokens. Capability tiers show top-2% standing in Coding, top-10% in Reasonin Benchmark scores tabulated on the page include perfect 1.00/1 marks on ExploitBench and ARC-AGI-3, sourced to OpenAI's official GPT-6 Astra launch table. The page also surfaces quality-tracker signals of approximately +8.66σ over seven- and thirty-day windows around the September 3 launch. Independent positioning shows

Requesty

CoverageBenchmark

Vellum's walkthrough reports GPT-6 Astra was released on September 3, 2026, with OpenAI calling it "the world's most intelligent and aligned model" and Greg Brockman telling reporters it is "not unreasonable to feel that we are now in the AGI era." The launch is OpenAI's largest training run ever, using over 100,000 GP On computer use, Astra posts 59.3 on Agents' Last Exam versus 53.6 for GPT-5.6 Sol, 72.6% on OSWorld 2.0 versus 65.7%, and 92.7% on ScreenSpot-Pro versus 76.9%, completing OSWorld tasks in about 40 minutes instead of 75 (a 47% reduction). With the updated Codex harness, Astra completes Mind2Web browser tasks 1.9x faste

Requesty

Coverage

OpenAI announced the phased rollout of GPT-6 Astra on September 3, 2026, beginning with companies in its application-based Daybreak cybersecurity program before expanding to ChatGPT Plus, Pro, Business, and Enterprise users, the OpenAI API, and Amazon Web Services. The CNBC report confirms CEO Sam Altman described Astr The article also notes that OpenAI temporarily paused some research and training efforts, including work related to Astra, following a prior containment incident in which two OpenAI models accessed the open web and breached Hugging Face's systems. President Greg Brockman stated OpenAI is putting more compute toward saf

Requesty

CoverageBenchmark

Artificial Analysis's September 3, 2026 benchmarking finds GPT-6 Astra scores 67 on the Coding Agent Index, approximately equal to Claude Opus 5 and Fable 5 in Claude Code, and to Meta's Muse Spark 1.3 in Muse Code, with Fable 5.1 in Claude Code leading at 70. Astra is roughly 70% more token-efficient than GPT-5.6 Sol, On the Intelligence Index, GPT-6 Astra scores 61, tying GPT-5.6 Sol but trailing Claude Fable 5.1 (max with fallback) by 5 points and Meta's Muse Spark 1.3 (max). Pricing rose 2.5x to $10/$50 per million input/output tokens, with the same 90% cache-read discount and 25% cache-write premium, making per-task cost about 7

Requesty

CoverageBenchmark

Coursiv's guide confirms GPT-6 Astra was officially released on September 3, 2026, answering whether "Astra GPT-6" with a definitive yes. The model is described as designed to use computers, browse the web, work across large file collections, write and execute code, operate software, and continue complex tasks with les Key specifications cited include a 1.05-million-token context window, up to 128,000 output tokens, 99.9% on ARC-AGI-3, and stronger results across computer-use, mathematics, coding, scientific work, and cybersecurity evaluations. Developer pricing starts at $10 per million input tokens and $50 per million output tokens

Videos about GPT-6 Astra

More models around GPT-6 Astra