Compare Models
GLM 5.3

Release Date: Aug 18, 2026

Developer zAI 🇨🇳
Context Window 1M
Max Output Tokens 128k
Token Costs (in/out) $1.40/4.40
Weights Open
Input Modalities

Accuracy

56.97 % ± 1.35

Cost / Test (Vals Index)

$ 6.732

Latency

66 min 29 s

Vals Index
BenchmarksAccuracyRankings
Vals Index

0.0%

±1.35
19/57
Code Migration

0.0%

±4.29
14/59
EMB

0.0%

±3.32
30/57
Finance Agent (v2)

0.0%

±2.07
13/60
Legal Research Bench

0.0%

±3.48
5/60
MedCode

0.0%

±2.11
38/92
MedScribe

0.0%

±2.00
6/94
ProofBench v1.1

0.0%

±5.02
16/31
Public Benefits Bench v1.1

0.0%

±1.21
4/35
Tax Agent Bench

0.0%

±2.97
3/20
TaxEval v2

0.0%

±0.88
66/145
Vibe Code Bench v1.1

0.0%

±3.94
15/95
Harvey's Legal Agent Benchmark

0.0%

±2.00
15/61
GPQA Diamond

0.0%

±1.85
36/138
IOI

0.0%

±7.48
7/27
LiveCodeBench

0.0%

±1.06
68/143
LegalBench

0.0%

±0.40
26/144
MMLU Pro

0.0%

±0.34
38/138
ProgramBench

0.0%

±0.86
7/45
SkillsBench

0.0%

±4.52
25/35
SWE-bench

0.0%

±0.94
6/88
Terminal-Bench 2.1

0.0%

±1.98
20/65
Terminal-Bench Science

0.0%

±2.44
16/25
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Fireworks AI
Temperature: 1
Top P: 0.95
Top K: Default
Max Output Tokens: 128,000
Reasoning Effort: max