ModelsNov 18, 2026 · 02:078 sources

OpenAI's GPT-6.2 Astra Scores First 100% on ARC-AGI-3 Amid Mixed Tester Reviews

From marketbrief.now

Image: marketbrief.now
AI;DR

OpenAI's newly launched GPT-6.2 Astra became the first AI model to achieve a verified 100% score on the ARC-AGI-3 reasoning benchmark, reaching the mark at a cost of $4,100 using a provider adapter while scoring 70.3% at $24,200 in its standard configuration. Astra also set state-of-the-art marks of 95.8% on ARC-AGI-2 at $0.50 per task and 98.5% on ARC-AGI-1 at $0.17 per task, prompting ARC Prize co-founder Greg Kamradt to state that static task benchmarks from GPT-1 through GPT-6.2 will no longer hold up for measuring future frontier models.

Despite the reasoning milestone and strong science scores, early testers reported mixed results in everyday use following the rollout of GPT-6.2 Astra and GPT-6.2 Sol. Wharton professor Ethan Mollick wrote that the launch is "the first time since GPT 4.5 that a frontier release from OpenAi is not a clear step forward," adding that the model was "often frustrating to use." Every CEO Dan Shipper called the models "a disappointing release compared to Fable 5.2" after weeks of testing, noting that Astra 6.2 was a step back in writing due to overly concise responses, frequently stopped too early on end-to-end coding builds, and struggled to vary i

Read the sourceSign in to suggest

Sources

+ 6 more
  1. SUPPORT@lyraxana“Artificial Analysis Intelligence Index Scores: 6.2 Astra @ max effort scores 62.1 6.2 Sol @ max effort scores 57.3 Fable 5.2 @ max effort scores 64.6”x.com
  2. SUPPORT@emollick“This is the first time since GPT 4.5 that a frontier release from OpenAi is not a clear step forward.”x.com
  3. SUPPORT@kimmonismus“GPT-6.2 Astra Max falls short 3 point behind Fable 5.2 in @ArtificialAnlys Intelligence Benchmark.”x.com
  4. SUPPORT@himanshustwts“GPT-6.2 Astra is flat with GPT-6 Astra, and GPT-6.2 Sol loses 3 points relative to GPT-6.1 Sol in the Artificial Analysis Coding Agent Index (Codex harness at max effort).”x.com
  5. SUPPORT@danshipper“We've been testing them for the last few weeks @every and, unfortunately, these models aren't great. Overall a disappointing release compared to Fable 5.2.”x.com
  6. SUPPORT@gregkamradt“GPT-6.2 Astra is the first model to achieve a 100% score on ARC-AGI-3.”x.com

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today