AI Index

GLM-4.6V-Flash

Language Zhipu AI Verified Released Dec 8, 2025
  • In TextImageVideo Out Text
  • Reasoning
  • Reasoning can be off
  • Tool use
  • Open weights

Prices

2 offerings · prices as listed
Z.ai API Official Global · USD GLM-4.6V-Flash
Conditions/ 1M tokensInputCache readOutput
Standard FreeFreeFree

Prices from Pricing (checked Nov 1, 2026)

Z.ai API Official China · CNY GLM-4.6V-Flash
Conditions/ 1M tokensInputCache readOutput
Standard FreeFreeFree

Prices from 智谱 BigModel：API 定价 (checked Nov 2, 2026)

Standard: the price a request with no conditions pays (the headline price elsewhere on the site).

Details

Family
glm-flash
Model ID

Not stated on Zhipu AI’s pages: max input, reasoning efforts, max thinking budget, structured output, built-in tools, knowledge cutoff, licence, aliases.

Run locally

Weights: zai-org/GLM-4.6V-Flash
  • Parameters 10.3B
  • Dense
  • Native precision BF16
  • Layers 40
  • KV cache per token 40 KB
  • Config positions 128K (config.json max_position_embeddings)
  • API context 125K (maker’s API)
My hardware GeForce RTX 4090 24 GB · 1008 GB/s · 64 GB DDR5 dual channel · 89.6 GB/s
On your hardware · GeForce RTX 4090
Runs well Q8_0 (automatic) 8K context 45–67 tok/sOne GPU
WeightsKV cacheOverheadUsable
Needs
≈ 11.0 GB (weights 10.2 + KV cache 0.3 + overhead 0.5)
Usable
21.6 GB (90% of its memory)
Generation speed
45–67 tok/s
Longest context
125K (the model’s limit)
First token
about 0.33–0.50 s for a 2K prompt
PrecisionWeightsNeeds at 8KOn your hardwareLongest context
BF16 (native)maker’s file19.2 GB≈ 20.4Runs well 24–36 tok/s37K
Q8_010.2 GB≈ 11.0Runs well 45–67 tok/s125K
FP89.6 GB≈ 10.4Runs well 47–71 tok/s125K
Q6_K7.9 GB≈ 8.7Runs well 57–86 tok/s125K
Q5_K_M6.8 GB≈ 7.6Runs well 66–99 tok/s125K
Q5_06.6 GB≈ 7.4Runs well 68–102 tok/s125K
Q4_K_M5.8 GB≈ 6.6Runs well 76–115 tok/s125K
Q4_05.4 GB≈ 6.2Runs well 82–123 tok/s125K
IQ4_XS5.1 GB≈ 5.9Runs well 87–130 tok/s125K
Q3_K_Mquality drops4.7 GB≈ 5.5Runs well 94–141 tok/s125K
IQ3_XXSquality drops3.7 GB≈ 4.5Runs well 118–177 tok/s125K
Q2_Kquality drops3.6 GB≈ 4.4Runs well 120–180 tok/s125K
IQ2_XXSquality drops2.5 GB≈ 3.3Runs well 169–253 tok/s125K

Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.

Compared with other hardware

Each at the highest precision it runs. Click one for details.

Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.

Run commands

The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.

Memory to fine-tune
Full≈ 155.9 GBDoesn’t fit your hardware
LoRA≈ 22.0 GBDoesn’t fit your hardware
QLoRA≈ 8.2 GBFits your hardware

Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.

Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated

Cited scores

From third-party leaderboards, not measured by us

This model isn’t on any leaderboard we cite yet.

Changes

All changes →
  1. Now priced on Z.ai API, China Input Free · Output Free · Cache read Free

Sources

Zhipu AI’s official policies: data use, privacy, terms →

Zhipu AI · China · z.ai

Snapshots, the same family and other models from this provider