Model
Model explorer

GLM-4-9B-Chat

OPEN
Zhipu AI · GLM-4 family · released Jun 5, 2024

Open 9B GLM-4 checkpoint (co-developed by Zhipu AI/THUDM Tsinghua) beating Llama-3-8B on semantics/math/code/reasoning; 26 languages, 128K context (separate 1M variant exists).

ReasoningCodingVisionFunction callingTool useAgentic
629.9
Elo · rank #278
Parameters
9B
Active params
9B (dense)
Context
128K tokens
Architecture
Dense transformer (GLM-4 architecture)
License
GLM-4 License (custom, free for research & most commercial use)
Languages
26+
API price (in/out)
No hosted API
Modalities
text
Benchmark results
Bar shows position within the tracked field; marker = field best
BFCLAgents81.0%#10
best: SmolLM3 3B (No Thinking) · 92.3%
C-EvalKnowledge75.6%#37
best: Qwen3.6-Plus · 93.3%
GSM8KMath79.6%#84
best: Llama 3.1 405B · 96.8%
HumanEvalCoding71.8%#79
best: Claude Opus 4.5 · 99.4%
IFEvalReasoning69.0%#115
best: Gemma 4 26B A4B · 98.5%
LogicKorHuman preference6.96#19
best: GPT-4o · 9.33
MATH-500Math50.6%#129
best: GPT-5 · 99.4%
MGSMMath65.3%#37
best: OpenAI o4-mini · 93.7%
MMLUKnowledge72.4%#134
best: OpenAI o3 · 92.9%
MMMLUKnowledge56.6%#51
best: Gemini 3.1 Pro · 92.6%
MT-BenchHuman preference8.35#35
best: Hunyuan-Large (A52B) · 9.4
XWinogradReasoning73.1%#5
best: Sarashina2-70B · 91.8%
Run it locally
VRAM @ Q4
6 GB
VRAM @ FP16
18 GB
Fits on (Q4)
RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB
Throughput data unavailable.
Quantizations
GGUF Q4 · GGUF Q8 · MLX
GLM-4 family
Elo progression across releases
GLM-4-9B-ChatJun 2024629.9GLM-4V-9BJun 2024675.7GLM-4-PlusAug 20241115.1
Resources
HFzai-org/glm-4-9b-chatGHzai-org/GLM-4APIAvailable via Zhipu AI API
API price weights · each benchmark row carries its own source badge (see methodology)