Skip to content
Leaderboard / Z.ai

GLM-5.3

Z.ai Max OPEN MIT
Verdict

GLM-5.3 is a strong all-rounder: #49 of 406 by SIscope score. Strongest at following instructions, weaker at agents & tools. Mid-range on price. Weights are open, so you can run it yourself.

SIscope score
79
#49 of 406
Human-vote rank
#28
of 419 · score 1474
Price per 1M tokens, blended
$0.96
$0.08 in · $3.6 out
Context window
1M
1,048,576 tokens

Shape

Each spoke is a category, 0 at the centre, 100 at the edge.

CodingAgentsReasoningMathChatInstructions

Category scores

Percentiles: 90 means better than about 90% of tested models.

Coding88
Agents & tools61
Reasoning & knowledge77
Math81
Chat quality94
Following instructions96

What goes into each category · Compare with Gemini 4 Carbon

Human-vote score over time

Text arena, recent updates.

146514731480Oct 20: 1473.7 (rank #27)Oct 22: 1473.9 (rank #27)Oct 28: 1473.8 (rank #28)Oct 20Oct 28

Coding votes over time

WebDev arena, recent updates.

161516231630Oct 16: 1621.1 (rank #23)Oct 18: 1621.6 (rank #23)Oct 19: 1621.2 (rank #25)Oct 22: 1622.7 (rank #24)Oct 23: 1622.8 (rank #25)Oct 27: 1622.7 (rank #26)Oct 28: 1622.8 (rank #27)Oct 16Oct 28

Independent test runs

Run by Epoch AI on the same harness for every model. Bars show the share of questions answered correctly.

GPQA Diamond91%
PhD-level science multiple choice questions
Better than 88% of 333 tested models. Run Aug 24, 2026.
SimpleQA Verified41%
Short factual questions; rewards not making things up
Better than 49% of 100 tested models. Run Aug 28, 2026.
OTIS Mock AIME 2024-2591%
Competition maths problems in the style of AIME
Better than 74% of 313 tested models. Run Aug 24, 2026.
FrontierMath Tiers 1-369%
Very hard, unpublished research-style maths problems
Better than 75% of 122 tested models. Run Aug 25, 2026.

Human-vote arenas

People compare two anonymous answers and pick the better one. Scores work like chess ratings.

ArenaRankScoreRange (95%)VotesUpdated
Text arena
Everyday chat, blind human votes
#28 of 41914741469–147926,606Oct 28, 2026
WebDev arena
Building web apps, blind human votes
#27 of 15216231616–163011,150Oct 28, 2026
Vision arena
Questions about images, blind human votes
Not on this board yet
Agent arena
Real multi-step agent sessions
#23 of 571.2%0.4–2.0%96,216Nov 2, 2026

Source: Arena (formerly LMArena) leaderboard dataset, CC BY 4.0. Fetched Nov 4, 2026.

Price

Input
$0.08
per 1M tokens
Output
$3.6
per 1M tokens
Blended
$0.96
3 in : 1 out, per 1M
Context window
1M
max tokens per request
Speed
Coming soon
our own tests

List price from the OpenRouter models API (z-ai/glm-5.3), Nov 4, 2026. Providers can charge differently.