GLM-4.7-Flash model card (Hugging Face) Jan 19, 2026
- AIME 25 91.6
- GPQA 75.2
- LCB v6 64
- HLE 14.4
- SWE-bench Verified 59.2
- τ²-Bench 79.5
- BrowseComp 42.8
glm-4.7-flash
glm-4.7-flash.
29
Benchmarks with a score
0
Ranked first, out of five or more models
0
In the top three
Z.ai Jan 2026
glm-4.7-flash
Open weights Small
138.6
#188 of 404
Self-Reported Index, from the lab's own numbers. No Epoch Capabilities Index score yet.
- GPQA Diamond 60.5% #137 /229
- Political Science LLM Benchmark 60.2% #41 /42
- SusVibes 9.70% #8 /10
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 | |
|---|---|---|---|---|
| LMArena (Elo) | 1,364 | #164 of 335 | Official leaderboard | |
| Vectara Hallucination Leaderboard (2025 dataset) | 9.30% | #51 of 104 | Official leaderboard | |
| Judgemark v2.1 | 66.2 | #26 of 42 | Official leaderboard | |
| EQ-Bench 3 | 1,118 | #50 of 78 | Official leaderboard | |
| Longform Creative Writing (EQ-Bench) | 47.8 | #88 of 119 | Official leaderboard | |
| Wer wird Millionär? LLM Benchmark | 191.0 | #61 of 77 | Official leaderboard | |
| Creative Writing v3 | 55.2 | #96 of 120 | Official leaderboard | |
| Humanity's Last Exam | 14.4% | #125 of 154 | Lab model card | |
| SEA-HELM | 53.8 | #58 of 71 | Official leaderboard | |
| StudyArena Text | 1,174 | #77 of 85 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| AIME 2025 | 91.6% | #43 of 143 | Lab model card | |
| OTIS Mock AIME 2024-2025 | 58.3% | #126 of 212 | Epoch AI run | |
| GPQA Diamond | 60.5% | #137 of 229 | Epoch AI run |
Coding
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| DesignArena UI Component | 1,199 | #72 of 139 | Official leaderboard | |
| DesignArena Website | 1,187 | #80 of 151 | Official leaderboard | |
| DesignArena Code | 1,168 | #84 of 143 | Official leaderboard | |
| DesignArena Game Dev | 1,125 | #96 of 143 | Official leaderboard | |
| DesignArena 3D Design | 1,103 | #93 of 136 | Official leaderboard | |
| SWE-bench Verified | 59.2% | #114 of 155 | Lab model card | |
| LiveCodeBench v6 | 64% | #71 of 90 | Lab model card | |
| DesignArena SVG | 1,028 | #85 of 103 | Official leaderboard | |
| DesignArena Data Visualization | 1,121 | #100 of 118 | Official leaderboard |
Agents and tools
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| tau2-bench | 79.5% | #39 of 79 | Lab model card | |
| SnakeBench | 18 | #153 of 188 | Official leaderboard | |
| LLM Chess | -328 | #104 of 123 | Official leaderboard | |
| BrowseComp | 42.8% | #70 of 81 | Lab model card |
Safety and security
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| SusVibes | 9.70% | #8 of 10 | Official leaderboard | |
| nocot-bench (NCRI) | 55.9 | #153 of 185 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| Political Science LLM Benchmark | 60.2% | #41 of 42 | Official leaderboard |
Its lab's numbers put it at 138.6, number 188 of 404
The Self-Reported Index fits one score to the 7 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 136.3 to 141.0. How we compute it.
- AIME 2025 91.6
- tau2-bench 79.5
- GPQA Diamond 75.2
- LiveCodeBench v6 64
- SWE-bench Verified 59.2
- BrowseComp 42.8
- Humanity's Last Exam 14.4
Its lab's number next to a leaderboard
Where the benchmark's own leaderboard, or Epoch AI's run, also has this model. Gaps mostly come from different harnesses, effort levels and task subsets. Every pair.
| Benchmark | Lab says | Board says | Gap |
|---|---|---|---|
| GPQA Diamond | 75.2% GPQA | 60.5% Epoch AI run | +15 pts |
Its launch post
The benchmarks the lab chose to headline, with the scores it gave.