OpenAI has released GPT-6.1 Sol, a model offering near-Astra intelligence for coding, computer use, and professional work. It is priced at one-fifth of Astra's standard API input and output token prices, aiming to make advanced AI capabilities more accessible.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
OpenAI has released GPT-6.1 Sol, which the company says has near-Astra intelligence and is aimed at coding, computer use and professional work, with standard API input and output token prices both at one-fifth of Astra's. But the release gives no benchmark scores, no absolute Astra pricing and no model specs, so the claim can currently only be treated as a vendor's own statement.
OpenAI released an AI model called GPT-6.1 Sol, which is the only part of this release that can be confirmed as fact. Around this model, the official release gives two core selling points: one is intelligence level, and the other is price.
On intelligence level, OpenAI says GPT-6.1 Sol has near-Astra intelligence; on use cases, OpenAI says it is aimed at coding, computer use and professional work. Both statements come from the vendor's own account and are source claims rather than independently verified conclusions.
On price, OpenAI says GPT-6.1 Sol's standard API input and output token prices are both one-fifth of Astra's. This ratio applies to both the input and output directions, and is the most specific quantitative information in the release.
Judging from the structure of the release's wording, Astra is used in this release as the comparison baseline for GPT-6.1 Sol's intelligence level and price. In other words, both of GPT-6.1 Sol's core claims are relative values, not absolute ones.
The problem is that the source does not give an absolute price figure for Astra. Since the comparison baseline has no price tag, "one-fifth" cannot be converted into a specific price for GPT-6.1 Sol. What readers get is only a proportional relationship, not a number that can be used for cost estimates.
Likewise, "near-Astra intelligence" is also a relative statement. The source provides no benchmark scores or quantitative intelligence metrics to support this claim, so the actual capability gap is unclear.
A product release would normally cover the model's basic specs and how to access it, but this source does not explain GPT-6.1 Sol's context length, modality support, available regions or launch timing.
The source also does not say whether the model has officially opened for use or through which channels it is provided. This means that even if readers accept the capability and price claims in the release, they cannot use them to judge whether, or when, they will be able to use this model.
The release also does not mention safety evaluations, limitations or regulatory-related content. These omissions do not affect the fact of the release itself, but they do affect judgment of the model's full picture.
What can be confirmed is: OpenAI released GPT-6.1 Sol. What can be recorded as source claims is: OpenAI says it has near-Astra intelligence, is aimed at coding and computer use and professional work, and has standard API input and output token prices both at one-fifth of Astra's.
What cannot be confirmed is whether these claims themselves hold up. The only source is the release on OpenAI's official blog, which is a vendor's own account with no third-party independent verification. In the absence of benchmark scores, absolute Astra pricing and model specs, these statements should remain at the level of source claims.
Therefore, the reasonable attitude toward this model is: track it as an announced release, rather than adopt it as a verified conclusion about capability and cost. Only if independent evaluations or Astra pricing disclosures appear later could the above relative statements be turned into comparable absolute judgments.
All information in this article comes from the release on OpenAI's official blog, which is a vendor's own account with no third-party independent verification. The release does not give Astra's absolute price, so "one-fifth" cannot be converted into a specific price; it provides no benchmark scores or quantitative intelligence metrics, so "near-Astra" lacks verifiable support; and it also does not explain context length, modality support, available regions, launch timing or safety evaluations. The judgment in this article that Astra serves as the comparison baseline is an inference based on the wording structure of the release.
The release of GPT-6.1 Sol itself is certain, but its two core selling points—"near-Astra intelligence" and "one-fifth the price"—are both relative statements, and the comparison baseline Astra's absolute pricing and quantitative capability metrics have not been disclosed. Until independent verification appears, this looks more like a release claim awaiting scrutiny than a conclusion that can be used directly for decisions.
Primary report
Primary source