BestLLMfor Your hardware. Your LLM. Your call.
The Local Copilot Kit APIOpen data Find my LLM
Model fiche

GLM 4.7 Flash

By Zhipu AI · China

Updated 2026-08-31

chat code reasoning moe multilingual
Parameters
31B
License
MIT
Context
125k
VRAM (Q4)
19 GB
Released
February 2026

Overview

GLM-4.7-Flash is Zhipu AI's MoE coding model with 31B total parameters and ~3B active, delivering the best code performance-per-VRAM in the 30B class under a fully open MIT license.

When to pick this model

  • Local coding assistant on a single RTX 3090/4090
  • Agentic coding workflows needing top 30B-class benchmarks
  • Projects requiring a fully unencumbered MIT license
  • Multilingual (Chinese/English) code and chat tasks

VRAM requirements by quantization

VRAM REQUIRED (GB)81216243248Q4_K_M19 GBQ5_K_M23 GBQ8_035 GBFP1662 GB
QuantizationVRAM required
Q4_K_M (recommended)19 GB
Q5_K_M23 GB
Q8_035 GB
FP16 (no quantization)62 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

In practice, GLM 4.7 Flash wants a 24 GB card at Q4_K_M (19 GB). Stepping up to Q8_0 nearly doubles the footprint to 35 GB, and unquantized FP16 weights take 62 GB — plan your GPU around the Q4 or Q5 figure unless you specifically need the higher fidelity.

Without a GPU, GLM 4.7 Flash needs roughly 33 GB of system RAM to run on CPU via llama.cpp or Ollama — workable for background jobs, but far slower than GPU inference. Throughput estimates from our compatibility engine: around 15 tokens/sec on entry-level GPUs, on the order of 40 tokens/sec on a mid-range card, and up to 100 tokens/sec on high-end hardware — assuming the chosen quantization fully fits in VRAM.

What hardware do you need

The table below matches GLM 4.7 Flash to common GPU memory tiers, using the highest-fidelity quantization that fully fits each card class. Spilling layers to system RAM works but costs most of the speed, so size your card to the quantization you actually want to run.

GPU memoryExample cardsBest fit for GLM 4.7 Flash
8 GBRTX 5070 Laptop, RTX 5060, RTX 5060 Ti 8GBDoes not fit — needs 19 GB at Q4_K_M
12 GBRTX 5070, RTX 5070 Ti Laptop, RTX 4080 LaptopDoes not fit — needs 19 GB at Q4_K_M
16 GBRTX 5080, RTX 4080 Super, Radeon RX 9070 XTDoes not fit — needs 19 GB at Q4_K_M
24 GBRTX 4090, Radeon RX 7900 XTX, RTX 5090 LaptopQ5_K_M (23 GB used)
32 GBRTX 5090Q5_K_M (23 GB used)

Which hardware should you buy to run GLM 4.7 Flash?

To run GLM 4.7 Flash locally at Q4, you need ~19 GB of VRAM. The best value for this today is a GMKtec EVO-X2 64GB (Ryzen AI Max+ 395 mini PC) (64 GB unified memory, half the price of an RTX 5090).

Check GMKtec EVO-X2 64GB (Ryzen AI Max+ 395 mini PC) price on Amazon →

As an Amazon Associate, BestLLMfor earns from qualifying purchases, at no extra cost to you. It does not influence our independent rankings.

Published benchmark scores

BenchmarkScore
SWE-Bench Verified59.2
LiveCodeBench v664

Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.

Strengths

  • SWE-bench Verified 59.2 — top of the 30B class
  • MIT license, fully unencumbered
  • ~3B active parameters — 60-100 tok/s on a 3090/4090
  • 128K context

Limitations

  • ~19GB at Q4 — tight fit on 16GB cards
  • Less fluent in French than Qwen or Mistral

Typical workloads

In our catalog grid, GLM 4.7 Flash is filed under Code (30B Class), Code Agents, Multilingual ZH/EN — the use cases where its size/quality trade-off makes the most sense. Its tags translate to concrete workloads: code generation and review (pair it with an editor integration like Continue.dev or Cline); multi-step reasoning and math-flavoured tasks; multilingual workloads.

The 125k-token context window covers long chats and mid-sized documents, though very large retrieval workloads will need chunking. The MIT license is permissive, so shipping it inside a commercial product raises no special legal questions.

Architecture & training

Architecture: MoE 31.2B total · ~3B active (glm4_moe_lite) · 128K context

Training: GLM 4.7 family from Zhipu AI / Z.ai. Flash variant tuned for code and agents — the strongest model in the 30B class at release.

Verdict

The strongest coding model in the 30B class right now, MIT-licensed and fast enough for real-time local use.

Quick start

ollama run glm-4.7-flash

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.

Similar models worth comparing

Frequently asked questions

How much VRAM does GLM 4.7 Flash need?

At the recommended Q4_K_M quantization, GLM 4.7 Flash needs about 19 GB of VRAM. Q8_0 takes 35 GB, and unquantized FP16 weights take 62 GB.

Can GLM 4.7 Flash run without a GPU?

Yes — with roughly 33 GB of system RAM it runs CPU-only through llama.cpp or Ollama. Expect a fraction of GPU speed, which is fine for background or batch jobs but slow for interactive chat.

What context window does GLM 4.7 Flash support?

GLM 4.7 Flash supports a 125k-token context window (128,000 tokens).

Can I use GLM 4.7 Flash commercially?

Yes. GLM 4.7 Flash is released under MIT, a permissive open-source license that allows commercial use, modification and redistribution.

How fast is GLM 4.7 Flash on consumer hardware?

Our compatibility engine estimates on the order of 40 tokens/sec on a mid-range GPU and up to 100 tokens/sec on high-end cards, assuming the quantization fully fits in VRAM.

Which quantization of GLM 4.7 Flash should I download first?

Start with Q4_K_M (19 GB) — the standard size/quality sweet spot. Step up to Q5_K_M or Q8_0 only if you have VRAM headroom. On a 24 GB card you can run up to Q5_K_M.

Tools

Is GLM 4.7 Flash the right pick for you?

Compute self-hosted ROI → Back to catalog