Primary use stated or inferred cautiously from official documentation.
O
OPENAI / gpt-6-astra
GPT‑6 Astra
Generalist frontier model oriented towards complex tasks, use of tools and highly demanding professional work.
01 / SUMMARY
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
02 / TECHNICAL SHEET
Limits and integration
| API ID | gpt-6-astra |
|---|---|
| Model type | Text and reasoning |
| Access model | Paid proprietary |
| License | Proprietary |
| Deployment | Hosted API |
| Release | September 3, 2026 |
| Knowledge cutoff | April 30, 2026 |
| Entrance | Text · Image |
| Exit | Text |
| Context window | 1,050,000 tokens |
| maximum output | 128,000 tokens |
| Reasoning | low, medium, high, xhigh, and max effort |
| Published tools | Features · Web search · File search · Code interpreter · Hosted shell · Computer use · MCP |
| Structured Outings | Supported |
| Batch processing | Compatible · 50% discount |
| Prompt cache | Supported |
| Fine-tuning | Not compatible |
| Verified platforms | OpenAI API |
Complex reasoning, research, code, documents, and agents with tools.
Higher price per token; requires minimum permissions and review for sensitive actions.
03 / CAPABILITIES
What it can do
Advanced reasoning
Breaks down complex objectives and maintains lengthy work plans.
Use of tools
Operate in flows with navigation, code and professional environments.
Multimodal work
Interpret text and visual information within the same task.
Cybersecurity
High capacity accompanied by reinforced safeguards.
Input modalities
TextImage
04 / PRICES
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | 10.00 USD / 1 M tokens | Standard API rate |
| Cached input | 1.00 USD / 1 M tokens | Reading reused prefixes |
| Cache write or storage | 12.50 USD / 1 M tokens | The condition varies by provider |
| Standard output | 50.00 USD / 1 M tokens | May include reasoning tokens |
| Batch input | 5.00 USD / 1 M tokens | Asynchronous processing |
| Batch output | 25.00 USD / 1 M tokens | Asynchronous processing |
Requests with more than 272,000 input tokens apply price multipliers; tools may add per-call charges.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
05 / EVALUATIONS
How to read the results
“
OpenAI presents it as its highest-capability model for complex end-to-end work. The published results are from the provider and should be validated using your own tasks.
| Benchmark | Result | Metric | Source |
|---|---|---|---|
| Terminal-Bench Science 0.1 | 64,6 % | Accuracy | View source ↗ |
| GPQA Diamond | 96,0 % | Accuracy | View source ↗ |
| BenchCAD | 95,9 % | Geometric overlap | View source ↗ |
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
06 / VERSIONS
Chronology
GPT‑6 Astra
Launch and publication of the security summary.
ANALYSIS
Related analysis
AIME 2024: What an Accuracy Percentage Across 30 Problems Measures—and Why Two Apparently Identical Results May Not Be Comparable
An AIME 2024 figure can look straightforward while summarizing very different protocols. This guide explains what the dataset represents, how results change with the number of attempts, answer-selection methods, tools, and evaluation harnesses, and what information to request before comparing models.
22 Sep 2026 ↗
BenchCAD: What Reconstructing an Executable Mechanical Part Demonstrates—and What It Does Not Prove About Industrial Design
BenchCAD 1.0 measures whether a system can generate, modify, or interpret CadQuery programs whose geometry can be executed and compared with a reference. It is a useful signal for parametric reconstruction, but it does not replace validation of tolerances, function, manufacturing, or integration into an engineering workflow.
22 Sep 2026 ↗
Retries, limits, and duplicates in AI APIs: how to prevent a failure from executing an action twice
A timeout does not prove that an API did not process a request. This guide proposes a design based on identifiable operations, bounded retries, and state reconciliation for integrating models and agents without duplicating emails, purchases, data changes, or tool calls.
22 Sep 2026 ↗
Contracts for AI Applications: Detecting Model, SDK, or Tool Changes Before Production
A response that appears correct may still violate the format, authorization, provenance, or business rule that supports an AI workflow. This guide explains how to turn those assumptions into verifiable contracts and use them to block regressions before production.
19 Sep 2026 ↗
GPT‑6 Astra raises agent capability—and the bar for security governance
OpenAI’s new model combines reasoning, computer use and a very large context window. Its reported gains matter, but they do not replace local testing or controls over permissions, data and irreversible actions.
03 Sep 2026 ↗07 / SOURCES