500xEngineer.

All postsSHORT2 min read

GPT-6.2 Sol vs Astra: 4x Cheaper per Token, and When to Pick Each

OpenAI's GPT-6.2 Sol costs a quarter of Astra's API price per token. Here are the prices, how big the benchmark gap really is, and when to use each.

GPT-6.2 Sol is 4x cheaper per token than GPT-6.2 Astra, and OpenAI says it lands within half a point of Astra on real code. OpenAI announced both models on November 17, 2026. If you run coding agents on the API, picking between them is mostly a question of which tasks deserve the bigger bill.

Video not loading? Watch the Short on YouTube

The prices

These are OpenAI's Standard API rates per million tokens, for requests up to 272K input tokens:

Model Input Cached input Output
GPT-6.2 Sol (gpt-6.2-sol) $2 $0.10 $10
GPT-6.2 Astra (gpt-6.2-astra) $8 $0.40 $40

Every line is a 4x gap. Sol keeps GPT-6.1 Sol's prices. Astra's input and output prices are 20% lower than GPT-6 Astra's, according to OpenAI.

How close the scores are

OpenAI says that on Cognition's FrontierCode 2.0, which checks whether changes to real codebases meet maintainers' standards for merging, Sol comes within 0.5 points of Astra at about one-fifth the cost, with both at max reasoning. That's OpenAI's number, on one benchmark.

Artificial Analysis tests models independently. On its Intelligence Index v5, as of November 17, GPT-6.2 Astra (Max) scores 62 and GPT-6.2 Sol (Max) scores 57. Its model pages rank Astra #3 of 91 and Sol #11 of 91. That's a 5-point gap on its scale, against a 4x gap in per-token price.

When to pick each

OpenAI's own advice on its Codex models page: choose Astra for demanding work, for example tracing a bug across services, fixing it and testing the change. Choose Sol for everyday coding, writing and work across several sources, like editing code or turning project notes into a plan. The API changelog puts it the same way: Astra for demanding reasoning and coding tasks, Sol for everyday coding and agent workflows.

In practice, I'd make Sol the default for everyday code edits and send the hard, multi-service bugs to Astra.

Try it

Both models are in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users (Enterprise and Edu once a workspace admin enables it), and in the OpenAI API. In the Codex CLI:

codex --model gpt-6.2-sol

Here's how you get paid

My opinion: I'd sell dev teams a model router. Sol runs everyday code edits by default, Astra takes the hard bugs, and you hand them a token and cost report built from their own logs, before and after. No income guaranteed.

Daily email

Get each breakdown in your inbox.

One email each evening, right after the Short goes live: the AI tool it covers, the full write-up with sources, and how to get paid for it. Free, and you can unsubscribe anytime.

Read past emails on Substack

Sources

All postsWatch on YouTube