/Tech18h ago

Planetary VC founder Ramez Naam argues GLM-5.2 matching Claude Fable 5 on a planning benchmark shows open-weight models are keeping pace

The open-weight model scored 9.0 compared to Claude's 9.1

97752012.5K

#1861

Original post

Ramez Naam@ramez#1992inTech

Open weight models continue to nip at the heels of proprietary models. The claims that frontier models are pulling away seem thin to me. And given the timing, hard to see how GLM-5.2 could have substantially benefited from distillation of Fable.

Kilo@kilocode

Everyone benchmarks GLM-5.2 against the frontier now. So we did too.

We pulled GLM-5.2's plan up against Claude Fable 5's, the plan that won our last frontier round. Same prompt, same task, same rubric.

Fable scored 9.1. GLM-5.2 scored 9.0.

7:52 AM · Jun 19, 2026 · 2K Views

Sentiment

Users express interest in GLM-5.2 nearly matching Claude Fable 5 on frontier planning benchmarks because they find the results noteworthy.

Pos

100.0%

Neg

0.0%

1 comments with sentiment.

Cluster Engagement

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

Posts from X

Most Activity

VIEWS10.6KBOOKMARKS20LIKES67RETWEETS4REPLIES8

Brian Roemmele@BrianRoemmele

Like I said Open Source Anthropic Mythos class AI in GLM-5.2!

We see the same.

Time to pick a different bogeyman for Anthropic, this is now in everybody’s hands.

Kilo@kilocode

Everyone benchmarks GLM-5.2 against the frontier now. So we did too.

We pulled GLM-5.2's plan up against Claude Fable 5's, the plan that won our last frontier round. Same prompt, same task, same rubric.

Fable scored 9.1. GLM-5.2 scored 9.0.

5h10.6K6720

GuruVerseX@GuruVerseX

@BrianRoemmele wild how fast mythos tier tech just went from secret sauce to open buffet

4h5

Lonely nerd.@lonelynerd00

@BrianRoemmele 👇

https://www.reddit.com/r/Cervantes_AI/comments/1uagqki/the_frontier_without_a_moat/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_butto

4h4

leon1969@Leon1969

@BrianRoemmele Tx for this. Interesting.

4h1