GLM-5.3-Flash: benchmarks, price and rank
GLM-5.3-Flash from Zhipu is 33rd of 123 rated AI models on Spezia's leaderboard, strongest on Toolathlon. It costs $0.15 in, $0.50 out per million tokens.
- 33rd of 123 rated models
- Rating 1,454
- First on 1 of 13 benchmarks
- Open weights
- Released Aug 20, 2026
- 1 million tokens of context
Where it is strongest
- 1st of 37 on Toolathlon
Where it falls behind
- 19th of 21 on FrontierSWE
- 18th of 20 on CursorBench
- 52nd of 66 on FrontierMath Tier 4
Price: $0.15 in, $0.50 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.
- APEX-Agents52.8%
Long tasks written by investment banking analysts, management consultants and corporate lawyers from their own work, done across files and apps in a realistic work setup.
29th of 56. Independent result from Epoch AI.
- FrontierSWE18.1%
Open-ended technical projects that take hours to tens of hours, from making systems faster to building large programs and applied machine learning research. Each run gets a 20-hour budget.
19th of 21. Independent result from Epoch AI.
- Humanity's Last Exam55.3%
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.
12th of 74. The maker's own number, from Zhipu AI, via LLM Stats.
- Toolathlon78.4%
Long tasks across many real apps and their tools, such as email, calendars, spreadsheets, code repositories and online stores.
1st of 37. The maker's own number, from Zhipu AI, via LLM Stats.
- CursorBench36.8%
Ambiguous coding tasks that span many files, taken from real sessions in the Cursor code editor.
18th of 20. Independent result from Epoch AI.
- DeepSWE 1.163.4%
Software engineering tasks run in the same simple agent setup for every model.
20th of 27. The maker's own number, from Zhipu AI, via LLM Stats.
- FrontierMath55.8%
Hundreds of new, unpublished math problems written by mathematicians, from hard undergraduate level up to research, with answers a computer can check.
49th of 76. Independent result from Epoch AI.
- CharXiv Reasoning89.4%
Questions about real charts from scientific papers that need reasoning across the chart, not just reading off a value.
7th of 45. The maker's own number, from Zhipu AI, via LLM Stats.
- LMArena text1,471
People put a question to two anonymous models and vote for the better answer.
24th of 107. Independent result from LMArena.
- LMArena WebDev1,609
People ask two anonymous models to build the same web page or app and vote for the better one.
25th of 98. Independent result from LMArena.
5 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.