Vals AI

Benchmarks

Models

Comparison

Vals Smith

App Reports

Government

News

Blogs

About

Vals AI

Benchmarks

Models

Comparison

Vals Smith

App Reports

Government

News

Blogs

About

Model

OCT 12, 2026

zAI

GLM Legal

πŸ‡¨πŸ‡³

OCT 07, 2026

Anthropic

Claude Haiku 5.5

πŸ‡ΊπŸ‡Έ

OCT 06, 2026

Mistral

Mistral Large 4

πŸ‡«πŸ‡·

SEP 30, 2026

Google

Gemini 4 Argon

πŸ‡ΊπŸ‡Έ

SEP 29, 2026

OpenAI

GPT-6.1 Sol

πŸ‡ΊπŸ‡Έ

SEP 28, 2026

Anthropic

Claude Sonnet 5.5

πŸ‡ΊπŸ‡Έ

SEP 28, 2026

StepFun

Step 5 Preview

πŸ‡¨πŸ‡³

SEP 22, 2026

Anthropic

Claude Opus 5.5

πŸ‡ΊπŸ‡Έ

SEP 22, 2026

Anthropic

Claude Opus 5.5 (CVP)

πŸ‡ΊπŸ‡Έ

SEP 22, 2026

Fireworks

Ember-1

πŸ‡ΊπŸ‡Έ

SEP 22, 2026

OpenAI

GPT-6 Luna

πŸ‡ΊπŸ‡Έ

SEP 22, 2026

OpenAI

GPT-6 Sol

πŸ‡ΊπŸ‡Έ

SEP 21, 2026

SpaceXAI

Grok 4.7

πŸ‡ΊπŸ‡Έ

SEP 21, 2026

Xiaomi

MiMo V2.6 Flash

πŸ‡¨πŸ‡³

SEP 21, 2026

Xiaomi

MiMo V2.6 Pro

πŸ‡¨πŸ‡³

SEP 18, 2026

SpaceXAI

Grok Voice Transcribe 2.0

πŸ‡ΊπŸ‡Έ

SEP 10, 2026

DeepSeek

DeepSeek V4.1 Flash

πŸ‡¨πŸ‡³

SEP 08, 2026

Inception

Mercury 2.5

πŸ‡ΊπŸ‡Έ

SEP 05, 2026

Meta

Muse Spark 1.3 Max

πŸ‡ΊπŸ‡Έ

SEP 03, 2026

OpenAI

GPT-6 Astra

πŸ‡ΊπŸ‡Έ

SEP 03, 2026

Ant Group

Ling 3.0 Flash Fin

πŸ‡¨πŸ‡³

SEP 02, 2026

Google

Gemini 3.8 Flash

πŸ‡ΊπŸ‡Έ

SEP 02, 2026

Meta

Muse Spark 1.3

πŸ‡ΊπŸ‡Έ

SEP 01, 2026

Anthropic

Claude Fable 5.1

πŸ‡ΊπŸ‡Έ

AUG 30, 2026

Meta

Muse Voice Transcribe

πŸ‡ΊπŸ‡Έ

AUG 28, 2026

Tencent

Hy4 Preview

πŸ‡¨πŸ‡³

AUG 26, 2026

zAI

GLM 5.3 Flash

πŸ‡¨πŸ‡³

AUG 18, 2026

zAI

GLM 5.3

πŸ‡¨πŸ‡³
Compare Models
zAI Vals

GLM Legal

GLM Legal is an open-weight model from zAI, released October 12, 2026. Its best result is #6 of 77 on Harvey's Legal Agent Benchmark.

Release Date: Oct 12, 2026

Developer zAI Β πŸ‡¨πŸ‡³
Context Window 1M
Max Output Tokens 128k
Token Costs (in/out) $0.07/0.25
Weights Open
Input Modalities Text input supportedImage input not supportedVideo input not supportedFile input not supported

Accuracy

49.08 %

Avg. Cost (In/Out)

$ 0.07 / $ 0.25

Latency

37 min 25 s

Vals Index
BenchmarksAccuracyRankings
Legal Research Bench

41.83%

Β±3.43
22/76
Harvey's Legal Agent Benchmark

19.17%

Β±3.25
6/77
LegalBench

86.23%

Β±0.41
14/150
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
View Hyperparameter Settings
Default Provider : zAI
Some benchmarks may use different provider and parameters. Please refer to the benchmark page for more information.
Temperature: 1
Top P: 0.95
Top K: Default
Max Output Tokens: 128,000
Reasoning Effort: max
Vals AI
Benchmarks Models Comparison Vals Smith App Reports
About Methodology News Blogs Government
Security Careers

Copyright Β© 2026 Vals AI. All rights reserved.

X (Twitter) LinkedIn