Unverified 'GPT Astra' Model Appears on OpenRouter With 1.05M Token Context, No OpenAI Confirmation
OpenRouter is listing a model called 'OpenAI GPT Astra Latest' with a 1.05 million token context window and $10/$50 per-million-token pricing. OpenAI has made no public announcement, and the listing's own description says it is an auto-redirecting alias rather than a fixed model.
What was published
A listing on OpenRouter's model catalog references "~openai/gpt-astra-latest," described as pointing to "the latest model in the OpenAI GPT Astra family." The listing specifies a 1,050,000-token (1.05M) context window, pricing of $10.00 per million input tokens and $50.00 per million output tokens, and multimodal input support covering text, images, and files, with text-only output.
OpenAI has not published any announcement, blog post, or documentation referencing a model called "GPT Astra." No independent benchmark data, system card, or training cutoff date has been disclosed. Everything currently known comes from the OpenRouter listing itself.
Why this listing is different from a standard release
The naming convention is notable. The tilde prefix ("~openai") and the "Latest" suffix indicate this is not a fixed, versioned checkpoint but a routing alias — an endpoint that automatically redirects to whatever model OpenRouter currently maps to the "GPT Astra" family. That means the underlying weights served under this identifier could change without a new version number, and the specs listed (context window, pricing) describe the alias's current configuration rather than a permanent, citable model release.
This pattern is common on OpenRouter for "latest" or "auto" style aliases tied to major labs, which are designed for developers who want to always hit the newest available checkpoint without updating their API calls. It does not, by itself, confirm that OpenAI has shipped a new distinct model named "Astra."
Specs as listed
- Context window: 1,050,000 tokens
- Input pricing: $10.00 per 1M tokens
- Output pricing: $50.00 per 1M tokens
- Modality: text + image + file input, text output
- Access: OpenRouter API, model ID "~openai/gpt-astra-latest"
- Benchmarks: none disclosed
- Parameter count: not disclosed
- Training cutoff: not disclosed
The pricing tier ($10/$50 per million tokens) sits well above OpenAI's publicly listed flagship rates for GPT-4o and GPT-4.1, and the 1.05M token context window exceeds any context length OpenAI has officially confirmed for a shipping model. Both figures should be treated as claims tied to this specific OpenRouter listing until corroborated by OpenAI directly.
What this means
Treat "GPT Astra" as unconfirmed until OpenAI issues an official statement. The listing could reflect a genuine internal or limited-access model that leaked into a third-party router before a public announcement, a mislabeled or placeholder entry, or an alias configuration that will point to different underlying models over time. The large context window and premium pricing are consistent with a frontier-tier model, but none of the core claims — capability, benchmark performance, or even the existence of a model actually named "Astra" — have been verified by OpenAI. Developers building on this endpoint today should expect the underlying model to change without notice, since "Latest" aliases are explicitly designed to do that.
Related Articles
Unbiased releases Pareto 26.9: 262K-context composite model at $2.50/$7.50 per 1M tokens
Unbiased has released Pareto 26.9, a multimodal composite model with a 262K-token context window, priced at $2.50 per 1M input tokens and $7.50 per 1M output tokens. It is listed on OpenRouter with a Oct 13, 2026 release date. A cheaper 1.0M-context preview, Pareto 26.10, is already listed alongside it.
ScaleDown launches Summarize, a 1M-context abstractive summarization model at $0.04 per 1M input tokens on OpenRouter
ScaleDown has released Summarize, an abstractive summarization model with a 1M-token context window, listed on OpenRouter on October 13, 2026. OpenRouter shows $0.04 per 1M input tokens (20% off the $0.05 list price) and $0 per 1M output tokens. Parameter count and benchmark scores have not been disclosed.
StepFun releases Step 5 Preview: 600B MoE with 1M context at $1/$2.70 per 1M tokens
StepFun has listed Step 5 Preview, a sparse Mixture-of-Experts model with 600B total and 27B active parameters and a 1.0M-token context window. It is priced at $1 input and $2.70 output per 1M tokens on OpenRouter. StepFun positions it as its flagship model for agentic work.
OpenBMB releases MiniCPM-V 4.7 1B, an on-device vision model with 4x/16x visual token compression
OpenBMB has released MiniCPM-V 4.7 1B, a 1B-parameter multimodal model for single-image, multi-image and video understanding on phones, tablets and local hardware. OpenBMB claims it substantially outperforms Qwen3.5 0.8B in overall score and inference efficiency in Instruct mode. Benchmark figures, context window and training cutoff are not given in the available model card text.
Comments
Loading...