GLM-4.6: benchmarks, price and rank
GLM-4.6 from Zhipu is 100th of 123 rated AI models on Spezia's leaderboard. It costs $0.50 in, $2 out per million tokens.
- 100th of 123 rated models
- Rating 777
- Rated on 6 benchmarks
- Open weights
- Released Sep 30, 2025
- 205,000 tokens of context
Where it is strongest
Not in the top tenth on any benchmark in use.
Where it falls behind
- 84th of 98 on LMArena WebDev
- 46th of 56 on BrowseComp
- 57th of 74 on Humanity's Last Exam
Price: $0.50 in, $2 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- BrowseComp45.1%
Questions whose answers are hidden deep on the web and take many searches and pages to track down.
46th of 56. The maker's own number, from Zhipu AI, via LLM Stats.
- Humanity's Last Exam17.2%
About 2,500 questions written by experts in more than a hundred fields to be as hard as they could make them, many needing a diagram or image.
57th of 74. The maker's own number, from Zhipu AI, via LLM Stats.
- LiveCodeBench v682.8%
Programming contest problems collected after models were trained, from LeetCode, AtCoder and Codeforces.
16th of 43. The maker's own number, from Zhipu AI, via LLM Stats.
- LMArena text1,440
People put a question to two anonymous models and vote for the better answer.
50th of 107. Independent result from LMArena.
- LMArena WebDev1,342
People ask two anonymous models to build the same web page or app and vote for the better one.
84th of 98. Independent result from LMArena.
- SWE-bench Verified68%
Real bug reports and feature requests from popular open-source Python projects. The model has to change the code so the project's hidden tests pass.
38th of 56. The maker's own number, from Zhipu AI, via LLM Stats.
- GPQA Diamond81%
Graduate-level questions in biology, physics and chemistry, written by PhDs so that searching the web does not give the answer away.
The maker's own number, from Zhipu AI, via LLM Stats.
1 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.