Try Delta
← All models

GLM

Zhipu AI · 2026 · Agentic reasoning · vision variants

Zhipu’s agentic line for software engineering, browser automation, and multi-turn tools. Flash and vision variants give Delta a local “hands on keyboard” model for long tasks.

Maker
Zhipu AI
Architecture
Agentic reasoning · vision variants
Fits from
Laptop · 12 GB+ RAM
Context
up to 131K
Runs
Local Intelligence
AgentsCodingReasoningVisionSoftware agentsBrowser automationTool use

Sizes

GLM 4.7 FlashQ4_K_M

requires 30 GB+ memory · 131K context · Workstation

18 GBWeights
GLM 4.6V FlashQ4_K_M

requires 12 GB+ memory · 131K context · Laptop

6.2 GBWeights
AutoGLM Phone 9BQ4_K_M

requires 12 GB+ memory · 66K context · Laptop

6.2 GBWeights

Memory at context length

Approximate working set for GLM 4.6V Flash as context grows.

4K
7.7 GB
8K
8.1 GB
16K
8.6 GB
33K
9.1 GB
66K
9.6 GB
131K
10 GB

Related families

Nemotron Nano 3

NVIDIA hybrid MoE for agents: reasoning, tools, and 1M-token context on your hardware.

GPT-OSS

OpenAI’s first open weights since GPT-2. Reasoning, agents, and tools you run at home.

Nemotron 3 Nano

NVIDIA’s compact agentic 4B: long-context assistants that stay on a laptop.