NOUMENOpen app
NOUMEN

Z.ai · Open weights

GLM 4.7 Flash

Added Jan 19, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

GLM-4.7-Flash is a lightweight 30B model optimized for coding and agentic tasks. Balances high performance with efficiency.

  • Reasoning
  • Tools
  • Structured output
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.07
Output
$0.40
Cache read
$0.01

Reference prices as published by the model provider. API example

Specifications

Context window
200K
Max output
128K
Average output (7 days)
86 tokens
Longer than 8% of 468 models

Side by side

GLM 4.7 Flash compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

GLM 4.7 FlashZ.aiThis modelTrinity Large ThinkingarceeRemove
ProviderZ.aiarcee
Discovery tierOpen weightsOpen weights
AddedJan 19, 2026Apr 1, 2026
Limits
Context window200K262.14KBest
Max output128KBest80K
Average output86 tokens327 tokens
Pricing, USD per 1M tokens
Input$0.07Best$0.25
Output$0.40Best$0.90
Cache read$0.01—
Performance
Throughput28.7 tok/s—
Time to first token2.5 s—
Request success (24h)100.0%—
Benchmark indexes
Intelligence Index10.610.8Best
Coding Index—25.8
Capabilities
ReasoningYesYes
ToolsYesYes
Structured outputYesNo
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • GLM 4.7 Flash
  • Trinity Large Thinking

Intelligence Index

  • GLM 4.7 Flash10.6
  • Trinity Large Thinking10.8

GLM 4.7 Flash: Better than 22% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 25.8

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 1.3%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 0.2%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 255 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 336 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 1.2%

  • AA-LCRv1.1

    Reasoning across long documents

    • GLM 4.7 Flash: 20.3%
    • Trinity Large Thinking: 38.0%Top

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • GLM 4.7 Flash: 5.0%
    • Trinity Large Thinking: 15.8%Top

  • GPQA Diamond

    Graduate-level science questions

    • GLM 4.7 Flash: 45.2%
    • Trinity Large Thinking: 75.2%Top

  • IFBench

    Following precise output instructions

    • GLM 4.7 Flash: 46.3%
    • Trinity Large Thinking: 56.3%Top

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 0.5%

  • SciCode

    Scientific computing in Python

    • GLM 4.7 Flash: —
    • Trinity Large Thinking: 40.6%

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • GLM 4.7 Flash: 3.8%
    • Trinity Large Thinking: 22.7%Top

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • GLM 4.7 Flash: 91.8%Top
    • Trinity Large Thinking: 90.1%

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • GLM 4.7 Flash: 20.3%
    • Trinity Large Thinking: 38.0%Top

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
28.7 tok/s
Faster than 20% of 253 models
Time to first token
2.5 s
Quicker than 64% of 253 models
Prompt cache hits
14.1%
About $0.0615 per 1M input tokens with caching
Request success (24h)
100.0%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 98.2%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "z-ai/glm-4.7-flash",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="z-ai/glm-4.7-flash",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Z.ai

Other Z.ai models, starting with the same family.