← Back to radar

Release-note comparison

Compare what actually changed.

Choose releases

Latest models, normalized.

The newest four are selected by default. Add another release to replace the oldest selected item, or remove one while keeping at least two in the comparison.

Field
Z.ai

GLM-5.2

16 Jun 2026

Z.ai

GLM-5.3

14 Aug 2026

LifecycleOpen weights since 16 Jun 2026Open weights since 14 Aug 2026
ImpactMajorMajor
Change typeWeights openedWeights opened
Capability shiftGeneral · Coding · Open weightsGeneral · Coding · Open weights
AvailabilityMIT open weightsAPI + open weights
Access statusStatus: Tracked · Channel: Weights · Weights: Open · Pricing: Unclear · Docs: SourceStatus: Tracked · Channel: API · Weights: Open · Pricing: Unclear · Docs: Source
Pricing signalNo pricing change found in the indexed note.No pricing change found in the indexed note.
Context signal1M context1M context
ConfidenceInferredConfirmed
Weights openedInferred

GLM-5.2

ImpactMajor
LifecycleOpen weights
Since 16 Jun 2026
Context1M contextAccessopenWeightsopenPriceunknownCheckInferred
Availability
StatusTrackedChannelWeightsWeightsOpenPricingUnclearDocsSource
Source qualityLaunch announcementFirst-party source · 1 receipt

Combines a stable million-token window with multiple reasoning-effort levels and openly released weights.

Weights openedConfirmed

GLM-5.3

ImpactMajor
LifecycleOpen weights
Since 14 Aug 2026
Context1M contextAccessopenWeightsopenPriceunknownCheckConfirmed
Availability
StatusTrackedChannelAPIWeightsOpenPricingUnclearDocsSource
Source qualityLaunch announcementVerified source · 1 receipt

Improves Z.ai’s coding benchmark by 50% over GLM-5.2 using scaled post-training.