GLM-5.2: benchmarks, price and rank
GLM-5.2 from Zhipu is 44th of 123 rated AI models on Spezia's leaderboard, strongest on IMO-AnswerBench. It costs $0.75 in, $2.40 out per million tokens.
- 44th of 123 rated models
- Rating 1,386
- Rated on 14 benchmarks
- Open weights
- Released Jun 16, 2026
- 1 million tokens of context
Where it is strongest
- 2nd of 18 on IMO-AnswerBench
Where it falls behind
- 41st of 56 on APEX-Agents
- 56th of 78 on SimpleQA Verified
- 46th of 69 on ARC-AGI-2
Price: $0.75 in, $2.40 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-Agents45.2%
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.
41st of 56. Independent result from Epoch AI.
- MCP Atlas76.8%
Multi-step tasks where the model has to pick and combine many real tools, connected through MCP, the standard way agents plug into apps and data.
9th of 33. The maker's own number, from Zhipu AI, via LLM Stats.
- SWE-Bench Pro62.1%
Long changes to real codebases in several languages, often across many files, including code from private and copyleft projects that models are unlikely to have seen.
16th of 50. The maker's own number, from Zhipu AI, via LLM Stats.
- Humanity's Last Exam54.7%
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.
13th of 74. The maker's own number, from Zhipu AI, via LLM Stats.
- Vending-Bench 2$8,314
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
15th of 60. Independent result from Epoch AI.
- Toolathlon48.2%
Long tasks across many real apps and their tools, such as email, calendars, spreadsheets, code repositories and online stores.
22nd of 37. The maker's own number, from Zhipu AI, via LLM Stats.
- DeepSWE 1.144%
Software engineering tasks run in the same simple agent setup for every model.
Independent result from LLM Stats.
- SimpleBench58.8%
Multiple-choice puzzles about everyday situations, space, time and people, written to trip up models that pattern-match instead of think. Ordinary adults do well on them.
44th of 68. Independent result from Epoch AI.
- SimpleQA Verified34.2%
Short factual questions with one checked answer, asked without web search.
56th of 78. Independent result from Epoch AI.
- FrontierMath59.2%
Hundreds of new, unpublished math problems written by mathematicians, from hard undergraduate level up to research, with answers a computer can check.
43rd of 76. Independent result from Epoch AI.
8 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.