Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage & billing
API
Suggestions and bugs
Use cases

|

Terms

|

Privacy

|

Refunds
Usage & billing
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksNano RPCEarn crypto
  1. Models
  2. Text
  3. Z.AI

GLM 4.6 Thinking

Thinking version of the latest GLM series chat model with strong general performance. Quantized at FP8

  • Reasoning
  • Tool Calling
  • Structured Output

Added Sep 29, 2025

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.55
Output
$2.20
Cache read
$0.11
Compare provider prices

Specifications

Context window
200K
Max output
65.5K
Parameters
357B / 32B
Total / active

Benchmarks

Sourced from Vectara.

Hallucination Rate

9.5%

Hallucination Rate: 9.5
Better than 51% of models compared

Factual Consistency

90.5%

Factual Consistency: 90.5
Better than 51% of models compared

Answer Rate

94.5%

Answer Rate: 94.5
Better than 13% of models compared

Avg Summary Length

Average generated summary length

77.2

Last updated 2026-10-22 · Matched as zai-org/GLM-4.6

Vectara Leaderboard

Leaderboard results

These leaderboards tested GLM 4.6, another version of this model.

BenchmarkScoreRank on NanoGPT
EQ-Bench

EQ-Bench data, MIT · Updated Nov 12, 2026

Creative Writing v3

1412

#68 of 96

Longform Writing

57.3

#61 of 97

UGI Leaderboard

UGI Leaderboard data, Apache-2.0 · Updated Oct 16, 2026

UGI Score

49.8

#42 of 149

UGI Willingness

4.2

#58 of 149

UGI Writing

53.3

#55 of 149

Filtered and ranked to models on NanoGPT.

Providers

Choose or restrict providers with the API. Provider-selection prices include the selection markup. Provider selection guide · Pricing API (JSON, USD per 1K tokens)

Auto
Auto (ZDR)
$0.55$2.20$0.11N/AN/A
Auto
Auto (Anonymized)
$0.55$2.20$0.11N/AN/A
MegaNova
MegaNova
US
$0.472$2.00N/AN/AN/A
DeepInfra
DeepInfra
FP4
US
$0.525$2.10$0.105N/AN/A
Novita
Novita
BF16
US
$0.578$2.31$0.116N/AN/A
Z.AI
Z.AI
FP4
SG
$0.63$2.31$0.116N/AN/A
Privacy:ZDR, prompts not retainedNo training on promptsMay train on prompts
Batch API
ProviderInputOutputCached input
Auto
Autoz-ai/glm-4.6:thinking
$0.55/M$2.20/M$0.11/M

Batch rows for this model use normal routing and cost the same as standard API requests. Set provider in a row to choose a provider.

Related text models

Compare GLM 4.6 Thinking with similar models from the same provider or model family.

GLM 5.3 FlashX Legal

z-ai/glm-5.3-flashx-legal

GLM 5.3 FlashX Legal is Z.AI's model trained for legal document analysis, contract review, legal reasoning and initial drafts. It accepts text and images and returns text, with always-on reasoning, tool calling, JSON output, a 1M-token context window and up to 131,072 output tokens. External legal databases and document-editing tools require your own integrations.

Apex Flash 1

z-ai/glm-5.3-flash-apex

Apex Flash 1 is Cantina Security and Yeta's cybersecurity-focused post-train of GLM 5.3 Flash for code investigation and tool use.

Apex Flash 1 Abliterated

z-ai/glm-5.3-flash-apex-abliterated

Apex Flash 1 Abliterated is an experimental derivative of Cantina Security and Yeta's Apex Flash 1 with modified refusal behavior.

GLM 5.3 Uncensored

z-ai/glm-5.3-uncensored

An experimental uncensored version of the full GLM 5.3 reasoning model for chat, creative writing, coding, and tool use. It may show unusual behavior or inconsistent response quality.

GLM 5.3 Flash Cybersecurity

z-ai/glm-5.3-flash-cybersecurity

GLM 5.3 Flash Cybersecurity is a cybersecurity-focused variant based on the uncensored model, with provider moderation for illegal activities. It supports always-on reasoning, image understanding, tool calling, and a 1,048,576-token context window.

GLM 5.3 Flash

z-ai/glm-5.3-flash

ox-alpha out of stealth! GLM-5.3 Flash is Z.ai's first natively multimodal GLM-5 model, with 320B total parameters and just 18B active parameters for efficient coding, agentic work, and precise 1M-token context. Its hybrid sparse-and-linear attention architecture helps it outperform GLM-5.2 at one-tenth the price while approaching Claude Opus 4.8 on coding and agentic benchmarks.