Microsoft launches its own coding model
MAI Code 2 and Code 2 Flash · in GitHub Copilot
Models
In ChatGPT, Codex, the API and GitHub Copilot
Written with AI assistance, reviewed by the ACN desk.
Why it matters
with top models level on price, developers will choose on coding results, and OpenAI's newest model didn't improve them.
OpenAI has released GPT-6.2 Astra and GPT-6.2 Sol, and its top model now costs the same as Anthropic’s. The Decoder reports that Astra’s API price is 20% below its predecessor’s. Astra costs $8 per million input tokens and $40 per million output tokens.
VentureBeat reports that Anthropic launched Claude Fable 5.2 eight days earlier at the same rates. VentureBeat says Anthropic charges half as much as OpenAI for cache reads. GPT-6.2 Sol keeps its predecessor’s price of $2 input and $10 output per million tokens, according to The Decoder.
Two independent rankings place Astra at or near Fable 5.2. Astra scores 174 on Epoch AI’s Capabilities Index, tying Fable 5.2, The Decoder reports. On the Artificial Analysis Intelligence Index, Astra finishes second, just behind Fable 5.2. On the Artificial Analysis Coding Agent Index, Astra stays level with GPT-6 Astra. Sol drops 3 points from GPT-6.1 Sol on the same index.
According to The Decoder, OpenAI designed Astra for specialized work in science, law and finance. OpenAI aims Sol at quick, everyday tasks. Both models are in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users. GitHub’s changelog says both are now generally available in GitHub Copilot. GitHub says they are designed for agentic coding and code review. It says GPT-6.2 Sol caught nearly twice as many real issues as GPT-6.1 Sol in early testing.
Sponsor
Spot an error? Email corrections@aicodingnews.si. Fixes are logged on the corrections page.
Sponsor
Sponsor
MAI Code 2 and Code 2 Flash · in GitHub Copilot
Astra's API price drops 20%
Mythos 5.2 for vetted security and science teams
The Riemann proof passes its formal check; understanding it comes next
Sponsor
Sponsor