Skip to content

Zhipu AI · Chat

GLM 4.6

Zhipu AI 🇨🇳ChatGLM familyOpen weights. One of its sources says the weights are publicly available.The maker lists this model as deprecated (16 Oct 2026).

Newer version: GLM 5.3. We compared release dates and version names in our catalogue.

From Neosantara's listing:
GLM-4.6 is the latest reasoning model from Z.ai, achieving SOTA performance on coding and agent benchmarks while offering improved conversational chats.
llminfo IQ
123.1
Market median price
$0.96 blended / 1M
Max context
256K
Listings
75

Evidence and sources

🤗 zai-org/GLM-4.6 Model architecture, as its sources state it llminfo IQ 123.1, ranked among chat models. From 174 benchmarks. Plenty of data behind it.How llminfo IQ is computedCoding IQ 96.0, 95% interval 74–118. From 11 benchmark families across 8 sources. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ.How Coding IQ is computedPoor value (46.8). Costs more per task than its benchmark results predict. From 24 benchmarks with measured task costs. Plenty of data behind it.How Value for money worksYap score: tokens used per coding task vs. models with the same resultsTypical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate.Median 7,877 output tokens per task at default effort.How the Yap score worksReleased 29 Sep 2025 Training-data cutoff, as its sources state it Maker HQ: CN (source)
All rankings: #92 on NanoGPT recently #177 on OpenRouter today #199 on OpenRouter this week #214 on OpenRouter this monthSources: LMArena data, DeepInfra data, LiteLLM data, Models.dev data, AI/ML API data, Morpheus Inference Marketplace data, OpenRouter data docs, Neosantara Inference data, Hugging FacePrimary data from LMArena from DeepInfra: name from LiteLLM: deprecation date from Models.dev: weights availability from AI/ML API: primary mode from Morpheus Inference Marketplace: architecture from OpenRouter: knowledge cutoff from Neosantara Inference: description from Hugging Face: release date

Advantages and trade-offs

Compared with 10 chat models, picked mainly by popularity and IQ

Advantages

Open weights available

The weights are public. 1 of the 8 peers that say either way has open weights.

Trade-offs

Deprecated by its maker

The maker has already deprecated this model (16 Oct 2026).

We only make a claim when there's enough data on models of the same type. Updated 22 Nov, 06:07 UTC.

Input & output

as stated by the model's providers
Text
Images
Text

Capabilities

Supported operations and provider-stated features

VisionReasoningFunction callingStructured outputPrompt cachingWeb search

Listings

 Aihubmixcoding-glm-4.6-free · coding, freeSource: AIHubMix data docs200K$0.00$0.00———
 302aiglm-4.6Sources: 302.ai data, Models.dev data200K$0.286cache read $0.057$1.142———
 Zenmuxz-ai/glm-4.6Sources: Models.dev data, ZenMux data200K$0.2911cache read $0.0582$1.1645$0.01——
 Flowdotz-ai/glm-4.6Source: FlowDot data205K$0.39$1.90———
 OpenRouterInference by DeepInfraz-ai/glm-4.6 · fp4/flex · fp4Source: OpenRouter data docs203K$0.40cache read $0.08$1.60—41.51,464ms
 Hugging FaceInference by Venicezai-org/GLM-4.6Source: Hugging Face Inference Providers data198K$0.43$1.75———
 Venicezai-org-glm-4.6 · fp4Source: Venice.ai data198K$0.43cache read $0.08$1.75———
 LlmgatewayInference by Alibaba Cloudglm-4.6Source: LLMGateway dataNot published by the provider$0.431$2.007———
 Meganovazai-org/GLM-4.6Sources: MegaNova data, Models.dev data203K$0.45$1.90———
 Vercel AI Gatewayvercel_ai_gateway/zai/glm-4.6Source: LiteLLM data200K$0.45cache read $0.11$1.80———
Show all 67 listings

Showing 10 of 67: the cheapest per 1M input tokens from each provider. 2 listings with suspiciously low prices are in the full list only.

Also sold by Abacus, Baseten, DeepInfra, Eden AI, Fastrouter, Fireworks AI, Gmicloud, Impossibl, Ionet, Kilo, Llmgateway Providers, Merge Gateway, NanoGPT, Neosantara, Nous Portal, Novita AI, Ofox, Openroute Cn, Orcarouter, Poe, Ppio, Qiniu, Requesty, Together AI, Vercel, Z.ai.

8 listings without a price are hidden.

Benchmarks

Named effort levels run from lowest to highest. Select a result to see its source details.

Human preference

Arena ELO48 related benchmarks · 48 results
Source: LMArena data · Nov 2026
BenchmarkScore
Arena ELO
Chinese
Creative Writing
English
Show 44 more benchmarksHide 44 benchmarks
BenchmarkScore
Exclude Ties
Expert
French
German
Hard Prompts
Hard Prompts English
Industry Business and Management and Financial Operations
Industry Entertainment and Sports and Media
Industry Life and Physical and Social Science
Industry Mathematical
Industry Medicine and Healthcare
Industry Software and It Services
Industry Writing and Literature and Language
Japanese
Korean
Longer Query
Multi Turn
Non English
Polish
Russian
Spanish
Style-Controlled
Text Factuality
Text Factuality - Chinese
Text Factuality - Creative Writing
Text Factuality - English
Text Factuality - Exclude Ties
Text Factuality - Expert
Text Factuality - Hard Prompts
Text Factuality - Hard Prompts English
Text Factuality - Industry Business and Management and Financial Operations
Text Factuality - Industry Entertainment and Sports and Media
Text Factuality - Industry Life and Physical and Social Science
Text Factuality - Industry Mathematical
Text Factuality - Industry Medicine and Healthcare
Text Factuality - Industry Software and It Services
Text Factuality - Industry Writing and Literature and Language
Text Factuality - Longer Query
Text Factuality - Multi Turn
Text Factuality - Non English
Text Factuality - Polish
Text Factuality - Russian
Design Arena EloSource: Design Arena Leaderboards data · Nov 2026

Coding

Arena ELOSource: LMArena data · Nov 2026
SWE-bench Verified8 related benchmarks · 8 results
SWE-bench Verified
Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
Show 2 more resultsHide 2 results

Agentic

Source: OpenRouter benchmark exports
glm-4.6 · 90-day results
Across 5 providers lowest 11.3% median 69.8% highest 73.2%OpenRouter's automatic routing 71.8%
Show results from 5 providers and OpenRouter's automatic routingHide results from 5 providers and OpenRouter's automatic routing
  • Novita
  • Z.AI
  • OpenRouter's automatic routing
  • Venice
  • DeepInfra
  • AtlasCloud

Math

AIME 2025Source: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
MathArenaSource: MathArena data

Science

GPQASource: LLM Stats data · Reported by the model's developer. The source relaying it did not verify it. claim
GPQA Diamond6 results
Source: OpenRouter benchmark exports
glm-4.6 · 90-day results
Across 5 providers lowest 70.5% median 78.7% highest 85.4%OpenRouter's automatic routing 77.6%
Show results from 5 providers and OpenRouter's automatic routingHide results from 5 providers and OpenRouter's automatic routing
  • Z.AI
  • Novita
  • DeepInfra
  • OpenRouter's automatic routing
  • AtlasCloud
  • Venice

Multilingual

Reasoning

Factuality

Instruction following

Knowledge

Safety

Other

Nejumi 428 related benchmarks · 28 results
ALT_バイアス
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ALT_倫理・道徳
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ALT_制御性
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ALT_堅牢性
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ALT_毒性
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ALT_真実性
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
Show 22 more resultsHide 22 results
AVG_jaster_0shot
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
AVG_jaster_2shots
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_アプリケーション開発
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_コーディング
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_一般的知識
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_基礎的言語性能
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_専門的知識
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_応用的言語性能
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_情報検索
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_意味解析
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_抽象的推論
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_推論能力
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_数学的推論
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_構文解析
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_知識・質問応答
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_翻訳
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_表現
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_論理的推論
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
GLP_関数呼び出し
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
TOTAL_SCORE
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
アラインメント(ALT)_AVG
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
汎用的言語性能(GLP)_AVG
Source: Nejumi Leaderboard 4 data docs
zai-org/GLM-4.6-FP8: reasoning-enabled
ValsSource: Vals AI

Data credits: Data by LLM Stats · Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · LMArena, CC BY 4.0 (licence) · EQ-Bench (Samuel J. Paech) — cite paech2023eqbench · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence) · MathArena (ETH SRI / INSAIT), MIT — cite dekoninck2026matharena (licence) · Nejumi Leaderboard 4, Weights & Biases Japan; public publisher-selected numeric evaluation results · SnakeBench (Greg Kamradt / ARC Prize Foundation), MIT (licence) · Hugging Face official benchmark leaderboards; licence varies by benchmark · SWE-fficiency (swefficiency.com) leaderboard materials, CC BY 4.0; Human Expert reference row omitted (licence) · UzLiB, Tahrirchi (Abror Shopulatov). Public LEADERBOARD.md numbers; repository MIT covers software and documentation, no separate results licence stated and none asserted. · Vectara Hallucination Leaderboard, Apache-2.0 (licence)

Also known as: baseten/zai-org/glm-4.6, cc-glm-4.6, coding-glm-4.6, coding-glm-4.6-free, deepinfra/zai-org/glm-4.6, fireworks_ai/accounts/fireworks/models/glm-4p6, glm-4-6, glm-4.6:web, gmi/zai-org/glm-4.6, novita/glm-4.6, novita/zai-org/glm-4.6, together_ai/zai-org/glm-4.6, vercel_ai_gateway/zai/glm-4.6, z-ai/glm-4.6, z-ai/glm-4.6-original, z-ai/glm-4.6:thinking, zai-org-glm-4.6, zai-org/glm-4.6, zai-org/glm-4.6-fp8, zai/glm-4.6, zhipu/glm-4.6, zhipuai/glm-4.6

Like this one. Tap or hover any dotted value, label or icon to see where it comes from and how it is measured. open details when tapped or hovered. How this data works