GLM-4-0414 model series README (GitHub) Apr 14, 2025
- AIME24 80.8
- AIME25 63.6
- Omni-MATH 68.4
- LCB (2408-2501) 59.1
- SysBench (ISR) 81.2
- IFEval 84.5
- ArenaHard 90.6
- BFCL (v3) 70.5
- SuperGPQA 52.6
- GPQA 66.1
GLM-Z1-32B-0414
GLM-Z1-32B-0414 ranks first on SysBench.
11
Benchmarks with a score
1
Ranked first, out of five or more models
1
In the top three
Z.ai Apr 2025
GLM-Z1-32B-0414
Open weights Mid-size
132.9
#251 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- SuperGPQA 52.6% #29 /35
- LiveCodeBench v5 59.1% #30 /47
- SysBench 81.2% #1 /6
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SysBench | 81.2% | #1 of 6 | Lab model card | |
| Arena-Hard-Auto | 90.6% | #10 of 39 | Lab model card | |
| IFEval | 84.5% | #86 of 135 | Lab model card | |
| SuperGPQA | 52.6% | #29 of 35 | Lab model card |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Misguided Attention | 53.5% | #16 of 45 | Official leaderboard | |
| AIME 2024 | 80.8% | #38 of 88 | Lab model card | |
| GPQA Diamond | 66.1% | #201 of 302 | Lab model card | |
| AIME 2025 | 63.6% | #121 of 143 | Lab model card | |
| Omni-MATH | 68.4% | #4 of 4 | Lab model card |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| LiveCodeBench v5 | 59.1% | #30 of 47 | Lab model card |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Berkeley Function Calling Leaderboard (BFCL) | 70.5% | #15 of 41 | Lab model card |
Its lab's numbers put it at 132.9, number 251 of 404
The Self-Reported Index fits one score to the 8 capability results Z.ai reported for this model, on the scale of the Epoch Capabilities Index. Labs choose which benchmarks to show and how to run them, so read it as the lab's claim, not an independent measurement. The 90% interval runs from 130.8 to 134.9. How we compute it.
- Arena-Hard-Auto 90.6
- IFEval 84.5
- AIME 2024 80.8
- Berkeley Function Calling Leaderboard (BFCL) 70.5
- GPQA Diamond 66.1
- AIME 2025 63.6
- LiveCodeBench v5 59.1
- SuperGPQA 52.6
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.