GLM-5.3
Zhipu AI (Z.ai)Aug 14, 2026N/A
NewOpen weightsSubscription first
Z.ai's newest flagship, launched through GLM Coding Plan subscriptions first (Lite $18 / Pro ~$75 / Max ~$160 tiers). Headline jumps: Terminal-Bench 3.0 at 28.3 (from 4.6) and 84.5% CyberGym for security workflows. Same 744B-class base as GLM-5.2 with extended agentic post-training.
- Status
- Subscription first
- License
- Open weights · Custom
- Context window
- 1M
- Max output
- 128K
- Input · $/M
- —
- Output · $/M
- —
- Cached input · $/M
- —
- Blended · $/M
- —
- Parameters
- 753B MoE lineage
- Inputs
- Text + images
Benchmarks
This model has no published scores on the benchmarks we track yet. Specs and pricing below the fold are still fully sourced.
Reasoning & knowledge
GPQA Diamond
no public score
MMLU-Pro
no public score
Humanity's Last Exam
no public score
ARC-AGI-2
no public score
Coding
SWE-bench Verified
no public score
SWE-bench Pro
no public score
LiveCodeBench
no public score
Agents & computer use
Terminal-Bench 2
no public score
OSWorld 2.0
no public score
BrowseComp
no public score
Math
AIME 2026
no public score
Known strengths
- Terminal-Bench 3.0 jump
- Security/CyberGym strength
- Cheap subscription entry
Caveats
- Per-token API + open weights were promised only weeks after the sub-only launch
- No official per-token price yet — analysts use GLM-5.2 rates (~$4.40/M out)