Inference
EN

GLM-5.3 vs Qwen3.8 Max

Change models
GLM-5.3Z.AI
Qwen3.8 MaxAlibaba
IndexAll categories, out of 100.
65.8 ±3.5 best
65.6 ±2.5
GapIndex points below the top.
Top
−0.2 · no clear gap
CoverageIndex weight that has results.
85%
95% best
SpeedOutput tokens per second.
86/s best
41/s
EloLMArena rating and rank.
1476
1485 best
ReleasedMonth of release.
2026-08
2026-08
Input / 1MUS dollars per 1M input tokens.
$0.07
—
Output / 1MUS dollars per 1M output tokens.
$1.40
—
Batch output / 1MBatch price per 1M output tokens.
$0.989
—
1K chats2K tokens in, 500 out, each.
$0.84
—
1K documents20K tokens in, 1K out, each.
$2.80
—
1K agent tasks200K tokens in, 20K out, each.
$41.98
—
ContextMaximum tokens in one request.
1.05M best
1M
AgenticMulti-step tasks with tools.
68.2 best
63.7
CodingCode writing and repair.
65.7 best
64.3
ReasoningLogic problems and puzzles.
64.5
66.7 best
MultimodalTasks with images and text.
—
67.2
KnowledgeFacts and expert knowledge.
59.3
62.6 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
66.8
71.4 best
MathMath problems.
63.1
67.6 best

Benchmarks

62 shared
Only GLM-5.3: 34Only Qwen3.8 Max: 51
GLM-5.3
Qwen3.8 Max
GPQA diamond · Epoch AIKnowledge
90.9%35 / 217
92.7% best25 / 217
OTIS Mock AIME 2024-2025Math
91.1%58 / 200
99.4% best17 / 200
LegalBenchLegal
84.8% best29 / 149
83.6%51 / 149
TaxEval v2Finance
72.4%62 / 143
75.6% best14 / 143
LiveCodeBenchCoding
80.5%67 / 142
87.9% best10 / 142
GPQA Diamond · Vals AIKnowledge
88.1%35 / 135
93.7% best6 / 135
MMLU ProKnowledge
86.8%38 / 135
88.6% best18 / 135
Vibe Code Bench v1.1Coding
78.1% best36 / 123
64.7%55 / 123
MedScribeHealthcare
88.8% best11 / 109
84.9%33 / 109
MedCodeHealthcare
42.9% best49 / 106
40.7%61 / 106
SimpleQA VerifiedKnowledge
41.0%51 / 91
45.8% best41 / 91
Chess PuzzlesReasoning
37.0%39 / 90
46.0% best26 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath
68.8%31 / 90
74.7% best24 / 90
Legal Research BenchLegal
49.0% best10 / 90
47.6%15 / 90
SWE-benchCoding
95.4% best5 / 87
85.6%17 / 87
Terminal-Bench 2.1 · BenchLMAgentic
88.2% best10 / 86
86.6%15 / 86
Tax Agent BenchFinance
43.3% best7 / 82
32.1%23 / 82
LiveBench Data AnalysisReasoning
70.2%56 / 80
78.4% best24 / 80
LiveBench ReasoningReasoning
85.8%35 / 80
88.2% best24 / 80
Mystery Game PuzzlesReasoning
33.0%28 / 80
38.0% best19 / 80