Compare Models

GLM 5.3 Flash

GLM 5.3 Flash is an open-weight model from zAI, released August 26, 2026. Its best result is #8 of 108 on MedScribe.

Release Date: Aug 26, 2026

Developer zAI  🇨🇳
Context Window 1M
Max Output Tokens 128k
Token Costs (in/out) $0.07/0.25
Weights Open
Input Modalities

Accuracy

49.66 %

Avg. Cost (In/Out)

$ 0.07 / $ 0.25

Latency

65 min 45 s

Vals Index
BenchmarksAccuracyRankings
Code Migration

20.52%

±4.21
49/75
CyberBench v1.1

67.86%

±5.57
25/45
EMB

55.93%

±3.56
45/72
Finance Agent (v2)

57.85%

±2.06
12/76
Legal Research Bench

45.19%

±3.46
15/75
MedScribe

88.94%

±1.91
8/108
ProofBench v1.1

21.00%

±4.09
38/48
SRE Bench

0.76%

±0.54
27/33
Tax Agent Bench

31.55%

±3.35
19/67
TaxEval v2

75.59%

±0.84
16/145
Vibe Code Bench 1-100

16.03%

±3.28
14/22
Vibe Code Bench v1.1

30.76%

±5.28
69/110
Harvey's Legal Agent Benchmark

6.67%

±1.65
28/76
GPQA Diamond

86.36%

±2.13
43/138
IOI

52.50%

±2.62
21/42
LiveCodeBench

80.51%

±1.05
69/143
LegalBench

83.93%

±0.43
43/149
MMLU Pro

86.06%

±0.34
49/138
MMMU Pro

86.01%

±0.83
23/93
ProgramBench

0.00%

±0.00
40/63
SkillsBench

40.16%

±4.43
30/35
SWE-bench

92.00%

±1.21
10/88
Terminal-Bench 2.1

62.92%

±2.25
38/75
Terminal-Bench Science

2.86%

±2.01
32/39
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : zAI
Temperature: 1
Top P: 0.95
Top K: Default
Max Output Tokens: 128,000
Reasoning Effort: max

Compared with

GLM 5.3 Flash vs Qwen 3.7 Max 14 shared benchmarks GLM 5.3 Flash vs Claude Sonnet 4.6 16 shared benchmarks GLM 5.3 Flash vs Kimi K2.6 16 shared benchmarks GLM 5.3 Flash vs Grok 4.5 21 shared benchmarks GLM 5.3 Flash vs MiMo V2.5 16 shared benchmarks GLM 5.3 Flash vs MiMo V2.5 Pro 15 shared benchmarks GLM 5.3 Flash vs DeepSeek V4 17 shared benchmarks GLM 5.3 Flash vs MiniMax-M3 19 shared benchmarks