admin/glm47-benchmarks
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Redesign Mermaid diagram with modern visual layout

Browse Source
This commit is contained in:
Gemini AI
2025-12-25 00:46:30 +04:00
Unverified
parent 5205c860ba
commit 40bee5e0f9
1 changed files with 106 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

137
README.md
View File

@@ -30,40 +30,115 @@ GLM-4.7 demonstrates competitive performance against the newest generation of fl
*Note: Best scores per category are highlighted in $\color{green}{\text{green}}$. Data sourced from [Z.ai Official Blog](https://z.ai/blog/glm-4.7).* *Note: Best scores per category are highlighted in $\color{green}{\text{green}}$. Data sourced from [Z.ai Official Blog](https://z.ai/blog/glm-4.7).*
```mermaid ```mermaid
graph TD flowchart TD
subgraph "2025 Flagship Benchmark Comparison" subgraph GLM [🚀 GLM-4.7 Dominance]
M[Math - AIME 25] --> G1{<b>GLM-4.7: 95.7%</b>} direction TB
M --> C1[Claude Sonnet 4.5: 87.0%] MATH[🧮 MATH<br/>AIME 25]
M --> G2[Gemini 3.0 Pro: 95.0%] CODING[💻 CODING<br/>LiveCodeBench v6]
M --> D1[DeepSeek-V3.2: 93.1%] SCIENCE[🔬 SCIENCE<br/>GPQA-Diamond]
M --> P1[GPT-5.1 High: 94.0%] LOGIC[🧠 LOGIC<br/>HLE w/Tools]
ENGINEERING[⚙️ ENGINEERING<br/>SWE-bench]
CO[Coding - LiveCode] --> G2_C{<b>GLM-4.7: 84.9%</b>} AGENTIC[🤖 AGENTIC<br/>τ²-Bench]
CO --> C2[Claude Sonnet 4.5: 64.0%]
CO --> D2[DeepSeek-V3.2: 83.3%] GLM_MATH[<b>GLM-4.7<br/>95.7%</b>]
CO --> P2[GPT-5.1 High: 87.0%] GLM_CODING[<b>GLM-4.7<br/>84.9%</b>]
CO --> G2_CO[Gemini 3.0 Pro: 90.7%] GLM_SCIENCE[<b>GLM-4.7<br/>85.7%</b>]
GLM_LOGIC[<b>GLM-4.7<br/>42.8%</b>]
S[Science - GPQA] --> G3{<b>GLM-4.7: 85.7%</b>} GLM_ENG[<b>GLM-4.7<br/>73.8%</b>]
S --> C3[Claude Sonnet 4.5: 83.4%] GLM_AGENT[<b>GLM-4.7<br/>87.4%</b>]
S --> D3[DeepSeek-V3.2: 82.4%]
S --> P3[GPT-5.1 High: 88.1%] MATH --> GLM_MATH
S --> G3_S[Gemini 3.0 Pro: 91.9%] CODING --> GLM_CODING
SCIENCE --> GLM_SCIENCE
L[Logic - HLE w/Tools] --> G4{<b>GLM-4.7: 42.8%</b>} LOGIC --> GLM_LOGIC
L --> C4[Claude Sonnet 4.5: 32.0%] ENGINEERING --> GLM_ENG
L --> D4[DeepSeek-V3.2: 40.8%] AGENTIC --> GLM_AGENT
L --> P4[GPT-5.1 High: 42.7%]
L --> G4_L[Gemini 3.0 Pro: 45.8%]
end end
classDef glmNode fill:#00c853,stroke:#1b5e20,stroke-width:3px,color:#ffffff,font-weight:bold,font-size:14px subgraph COMPETITORS [📊 Top Competitors]
classDef sonnetNode fill:#f1f8e9,stroke:#c5e1a5,stroke-width:1px,color:#558b2f direction LR
classDef budgetNode fill:#e3f2fd,stroke:#2196f3,stroke-width:1px,color:#0d47a1
subgraph GEM [💎 Gemini 3.0 Pro]
G_MATH[95.0%]
G_CODING[<b>90.7%</b>]
G_SCIENCE[<b>91.9%</b>]
G_LOGIC[<b>45.8%</b>]
G_ENG[76.2%]
G_AGENT[<b>90.7%</b>]
end
subgraph GPT [🔵 GPT-5.1 High]
P_MATH[94.0%]
P_CODING[87.0%]
P_SCIENCE[<b>88.1%</b>]
P_LOGIC[42.7%]
P_ENG[76.3%]
P_AGENT[82.7%]
end
subgraph DS [🟢 DeepSeek-V3.2]
D_MATH[93.1%]
D_CODING[83.3%]
D_SCIENCE[82.4%]
D_LOGIC[40.8%]
D_ENG[73.1%]
D_AGENT[85.3%]
end
subgraph CLAUDE [🟣 Claude Sonnet 4.5]
C_MATH[87.0%]
C_CODING[64.0%]
C_SCIENCE[83.4%]
C_LOGIC[32.0%]
C_ENG[<b>77.2%</b>]
C_AGENT[87.2%]
end
end
class G1,G2_C,G3,G4 glmNode GLM_MATH -.-> G_MATH
class C1,C2,C3,C4 sonnetNode GLM_MATH -.-> P_MATH
class D1,D2,D3,D4,G2,P1,P2,P3,P4,G2_CO,G3_S,G4_L budgetNode GLM_MATH -.-> D_MATH
GLM_MATH -.-> C_MATH
GLM_CODING -.-> G_CODING
GLM_CODING -.-> P_CODING
GLM_CODING -.-> D_CODING
GLM_CODING -.-> C_CODING
GLM_SCIENCE -.-> G_SCIENCE
GLM_SCIENCE -.-> P_SCIENCE
GLM_SCIENCE -.-> D_SCIENCE
GLM_SCIENCE -.-> C_SCIENCE
GLM_LOGIC -.-> G_LOGIC
GLM_LOGIC -.-> P_LOGIC
GLM_LOGIC -.-> D_LOGIC
GLM_LOGIC -.-> C_LOGIC
GLM_ENG -.-> G_ENG
GLM_ENG -.-> P_ENG
GLM_ENG -.-> D_ENG
GLM_ENG -.-> C_ENG
GLM_AGENT -.-> G_AGENT
GLM_AGENT -.-> P_AGENT
GLM_AGENT -.-> D_AGENT
GLM_AGENT -.-> C_AGENT
classDef glmNode fill:#00C853,stroke:#004D40,stroke-width:3px,color:#FFFFFF,font-weight:bold,font-size:13px,radius:8px
classDef geminiNode fill:#FFB74D,stroke:#E65100,stroke-width:2px,color:#FFFFFF,font-weight:bold,font-size:12px,radius:6px
classDef gptNode fill:#64B5F6,stroke:#1565C0,stroke-width:2px,color:#FFFFFF,font-weight:bold,font-size:12px,radius:6px
classDef deepseekNode fill:#4DB6AC,stroke:#00695C,stroke-width:2px,color:#FFFFFF,font-weight:bold,font-size:12px,radius:6px
classDef claudeNode fill:#AB47BC,stroke:#6A1B9A,stroke-width:2px,color:#FFFFFF,font-weight:bold,font-size:12px,radius:6px
classDef categoryNode fill:#37474F,stroke:#263238,stroke-width:2px,color:#ECEFF1,font-weight:bold,font-size:11px,radius:4px
classDef subgraphNode fill:#FAFAFA,stroke:#B0BEC5,stroke-width:2px,stroke-dasharray: 5 5
class GLM_MATH,GLM_CODING,GLM_SCIENCE,GLM_LOGIC,GLM_ENG,GLM_AGENT glmNode
class G_MATH,G_CODING,G_SCIENCE,G_LOGIC,G_ENG,G_AGENT geminiNode
class P_MATH,P_CODING,P_SCIENCE,P_LOGIC,P_ENG,P_AGENT gptNode
class D_MATH,D_CODING,D_SCIENCE,D_LOGIC,D_ENG,D_AGENT deepseekNode
class C_MATH,C_CODING,C_SCIENCE,C_LOGIC,C_ENG,C_AGENT claudeNode
class MATH,CODING,SCIENCE,LOGIC,ENGINEERING,AGENTIC categoryNode
``` ```
| Category | Benchmark | **GLM-4.7** | Claude Sonnet 4.5 | GPT-5.1 High | DeepSeek-V3.2 | Gemini 3.0 Pro | Source | | Category | Benchmark | **GLM-4.7** | Claude Sonnet 4.5 | GPT-5.1 High | DeepSeek-V3.2 | Gemini 3.0 Pro | Source |