SUBSCRIBE

The new best open model beat GPT and Claude in every test. Try finding one you can verify.

01THE CLAIM
"Meet Qwen3.8-Max, our most capable model to date. Next week, the open weights of Qwen3.8-Max will be released" [SOURCE ↗]

THE MOVE: SELF-MARKED, graded by the party that benefits from the grade

TRUE, BUT3 SOURCES · LIVE 2026-09-05
ALIBABA QWEN TEAM TRACK RECORD2 CLAIMS · 40/100 BS RATE →
02THE CHECK

WHAT SURVIVES CHECKING. Qwen3.8-Max is genuinely frontier-class. Number two open model on the independent Vals Index. Real achievement.

WHAT DOES NOT. the launch table where it beats GPT-5.6 and Claude Fable 5 traces entirely to Alibaba's own unpublished internal evals. Not on the public leaderboard. No methodology. And the open weights promised for next week had not shipped.

THE PATTERN TO REMEMBER. strong model, homemade crown. This is the oldest play in AI marketing and it works because nobody checks.

03SAY THIS IN THE MEETING
"Ask any model vendor one question: 'Which of these benchmark numbers appears on a leaderboard you don't run?' Then count the seconds of silence."

On August 3, 2026, Alibaba's Qwen team published its launch announcement: "Meet Qwen3.8-Max, our most capable model to date. Next week, the open weights of Qwen3.8-Max will be released." The launch materials included a benchmark table in which Qwen3.8-Max beats GPT-5.6 and Claude Fable 5, and the an

🔒 THE FULL AUTOPSY · FREE WITH AN ACCOUNT

You just read the free check. Sign in free, a code by email, no passwords, and the rest unlocks: the evidence trail, the steelman and the rebuttal, all 3 sources with quotes and screenshots, and our on-record call.

This story is a stable, citable object. If you can falsify a verdict, tell us. Corrections are loud here.