Claude Sonnet 4.5 Thinking vs GLM-5 Thinking

Change models
IndexAll categories, out of 100.
53.3 ±6.9 best
52.9 ±9.2
GapIndex points below the top.
Top
−0.4 · no clear gap
CoverageIndex weight that has results.
80% best
65%
SpeedOutput tokens per second.
—
—
EloLMArena rating and rank.
—
—
ReleasedMonth of release.
2025-09
—
Input / 1MUS dollars per 1M input tokens.
—
—
Output / 1MUS dollars per 1M output tokens.
—
—
1K chats2K tokens in, 500 out, each.
—
—
1K documents20K tokens in, 1K out, each.
—
—
1K agent tasks200K tokens in, 20K out, each.
—
—
ContextMaximum tokens in one request.
200K
—
AgenticMulti-step tasks with tools.
56.6 best
56.0
CodingCode writing and repair.
50.1
52.7 best
ReasoningLogic problems and puzzles.
—
—
MultimodalTasks with images and text.
57.7
—
KnowledgeFacts and expert knowledge.
52.9
52.9
MultilingualTasks in many languages.
60.2
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
53.4
54.9 best

Benchmarks

14 shared
Only Claude Sonnet 4.5 Thinking: 14
Claude Sonnet 4.5 Thinking
GLM-5 Thinking
84.1%39 / 149
84.1%39 / 149
TaxEval v2Finance
73.3% best50 / 143
70.0%95 / 143
73.0%82 / 142
81.9% best58 / 142
81.6%61 / 135
83.3% best56 / 135
MMLU ProKnowledge
87.4% best28 / 135
86.0%50 / 135
CorpFin v2Finance
62.0%48 / 131
62.9% best46 / 131
22.6%94 / 130
23.4% best93 / 130
AIMEMath
88.2%30 / 95
91.7% best23 / 95
MedQAHealthcare
94.7% best15 / 95
94.3%18 / 95
SWE-benchCoding
70.0%61 / 87
71.4% best57 / 87
41.6%29 / 66
49.4% best21 / 66
IOI v1Coding
18.3%25 / 61
22.0% best20 / 61
62.2% best19 / 53
52.5%44 / 53
54.5% best16 / 51
53.2%21 / 51