openai
OpenAI: GPT-6 Astra
openai/gpt-6-astra
Our take
AI-assisted analysis based on the sources below. Suitability is an editorial assessment; test it on your own workload.
OpenRouter's catalog describes openai/gpt-6-astra as an OpenAI flagship for demanding end-to-end work, with text, image, and file inputs producing text. The listing reports a very large context window, always-on reasoning with selectable effort, and support for tools plus structured outputs. Because this source pack contains only catalog data rather than an independent model card, those capabilities are best treated as catalog-reported claims to validate on real workloads.[1]
Strengths
- Catalog-reported multimodal input: text, images, and files, with text responses.[1]
- The listing advertises a very large context window and mandatory reasoning with multiple effort settings.[1]
- Tool calling, configurable tool choice, structured outputs, response formatting, seed, and verbosity controls are listed.[1]
Limitations
- Only an OpenRouter catalog entry was supplied; no independent model card was available in this source pack, so benchmark fields and claimed capabilities need workload validation.[1]
- Text is the only listed output modality; there is no catalog evidence of image, audio, or file generation.[1]
- The catalog marks reasoning as mandatory and lists no non-reasoning mode; it also does not report a knowledge cutoff.[1]
Where to use it
- Evaluation for deep research, advanced analysis, scientific work, or document drafting that may benefit from a large context window.[1]
- Software-engineering or agent-style workflows where tool use and structured outputs can be tested against project requirements.[1]
- Reviewing mixed text, image, and file inputs when text-only responses are acceptable.[1]
Sources
- [1] OpenRouter catalog · OpenAI: GPT-6 AstraSources checked: Nov 13, 2026
About this model
Description below is the original text supplied by OpenRouter, shown in the source language.
GPT-6 Astra is OpenAI's flagship model for demanding end-to-end work. It is suited for advanced analysis, software engineering, deep research, scientific work, and document creation, with particular strengths in long-horizon...
Specifications
- Model ID
- openai/gpt-6-astra
- Canonical ID
- openai/gpt-6-astra-20260903
- Creator
- openai
- Added to OpenRouter
- 04.09.2026
- Context length
- 1,050,000
- Modality
- text+image+file->text
- Input modalities
- file, image, text
- Output modalities
- text
- Token pricing
- USD per 1M tokens: in $10 / out $50
- Knowledge cutoff
- —
- Expiration date
- —
Tasks and capabilities
A grade appears only where the source supplies a matching measurement for that task, the model is active and the cohort holds at least 10 comparable models. The capability column lists facts the source declares — supported formats, reasoning and tool parameters — which are not measured quality.
| Task | Grade | Source-declared capability |
|---|---|---|
| General quality | SS — top 10% of the cohort | reasoning: mandatory · default medium · efforts: max, xhigh, high, medium, low · parameters: include_reasoning, reasoning, reasoning_effort |
| Coding | SS — top 10% of the cohort | — |
| Agents | SS — top 10% of the cohort | parameters: structured_outputs, tool_choice, tools |
| Writing | No measurement | output: text |
| Reasoning | No measurement | reasoning: mandatory · default medium · efforts: max, xhigh, high, medium, low · parameters: include_reasoning, reasoning, reasoning_effort |
| Document analysis | No measurement | input: file |
| Vision | No measurement | input: image |
| Image generation | No measurement | — |
| Audio | No measurement | — |
| Video | No measurement | — |
| Embeddings | No measurement | — |
Evaluations
| Source | Benchmark | Value | Grade | Data retrieved |
|---|---|---|---|---|
| Artificial Analysis | Intelligence index | 56.9 | SS — top 10% of the cohort | 13.11.2026 |
| Artificial Analysis | Coding index | 76.9 | SS — top 10% of the cohort | 13.11.2026 |
| Artificial Analysis | Agentic index | 51 | SS — top 10% of the cohort | 13.11.2026 |
How grades work
Each letter grade compares a model only against other models scored by the same source and metric in the current snapshot. S means the top 10%, A the next 20%, B the next 30%, C the rest. Models without a published measurement are shown as not rated — a supported format alone never implies measured quality.
- S — S — top 10% of the cohort
- A — A — next 20% of the cohort
- B — B — next 30% of the cohort
- C — C — below the B band
Providers
Optional live provider statistics from OpenRouter. Missing values are unknown, not zero.
| Provider | Context length | Price in /1M | Price out /1M | Latency p50 (30m, s) | Throughput p50 (30m, tok/s) | Uptime (30m) |
|---|---|---|---|---|---|---|
| OpenAI | 1,050,000 | $5 | $25 | — | — | 100% |
| Amazon Bedrock | 1,050,000 | $10 | $50 | — | — | 100% |
| Azure | 1,050,000 | $10 | $50 | — | — | 100% |
| OpenAI | 1,050,000 | $10 | $50 | — | — | 100% |
| Amazon Bedrock | 1,050,000 | $11 | $55 | — | — | — |
| Azure | 1,050,000 | $11 | $55 | — | — | — |
| OpenAI | 1,050,000 | $20 | $100 | — | — | — |
| OpenAI | 1,050,000 | $60 | $300 | — | — | — |
Provider data retrieved: 13.11.2026
Related articles
- OpenAI study links fading chain-of-thought transparency to capability and compute trade-offsAn analysis by OpenAI's safety team attributes the declining monitorability of frontier models like GPT-6 Astra to four training factors, framing the loss of visible reasoning as a trade-off between safety, capability, and cost.
- ExpertCrowd-SWE: Expert-Authored Benchmark Puts Seven Frontier Coding Agents 35 Points ApartA Toptal ExpertCrowd technical report introduces EC-SWE, 100 engineer-written software tasks graded like maintainer review; seven frontier models span 35 points, with Claude Opus 5.5 and GPT-6 Astra statistically tied at the top.