GLM-4V
Parameters
9B
Context Length
128K
Modality
Multimodal
Auxiliary Parameters
1.1B
Architecture
Dense
KV Cache Rate
640 KB/tok
License
MIT License
Release Date
15 Jan 2024
Knowledge Cutoff
-
API Pricing (per 1M)
Input: $0.60 · Output: $1.80
Hardware Sizing
Precision
Context
Workload Memory: 11.69 GB
INT4 · 8K context
Base Weights: 4.19 GB
KV Cache: 5.00 GB
Runtime & Graphs: 2.50 GB
Minimum Required Hardware
1x NVIDIA L4 24GB
11.7 / 21.6 GB usable
1x RTX 4060 Ti 16GB
11.7 / 14.4 GB usable
Apple M4 24GB
11.7 / 18.0 GB usable
Determine requirements for your specific workload:
Open in VRAM Calculator
Architecture Diagram
Evaluation Benchmarks
Rank
#143
| Benchmark | Score | Rank |
|---|---|---|
General Text | 1352 | 144 |
Rankings
Overall Rank
#143
Coding Rank
-
About GLM-4V
GLM-4V is a multimodal vision-language model developed by Z.ai for high-resolution image analysis, visual question answering, and video comprehension. It integrates a 4K-capable Vision Transformer with the GLM language backbone for cross-modal reasoning.
Technical Specifications
Attention
Attention Structure
Multi-Head Attention
Attention Heads
32
Key-Value Heads
32
Attention Head Dimension
128
Position Embedding
Absolute Position Embedding
Normalization
RMS Normalization
Dimensions
Auxiliary Parameters
1.1B
Hidden Dimension Size
4,096
Number of Layers
40
FFN Intermediate Size (Dense)
13,696
Tokenizer
Vocabulary Size
151,552
About GLM Family
General Language Models from Z.ai