LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GLM 5.3 vs Qwen3.5 397B A17B

Updated October 21, 2026. 52 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GLM 5.3 Replace Qwen3.5 397B A17B
GLM 5.3Z.ai
Coding
73.1
Overall
44.8
Output $/M
$1.40
52
Qwen3.5 397B A17BQwen
Coding
28.6
Overall
21.4
Output $/M
$2.34
0
vs boards won of 52 shared

Decision reading

Pick GLM 5.3 for coding: it scores 73.1 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B.

Qwen3.5 397B A17B is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Model AGLM 5.3Z.ai73.1/1002 independent benchmarks$0.36 per solved taskModel BQwen3.5 397B A17BQwen28.6/1001 independent benchmarkNo cost per solved task
Shared results
52
GLM 5.3 only
48
Qwen3.5 397B A17B only
13
Like-for-like categories
8 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GLM 5.3 73.1 from 2 benchmarks, Qwen3.5 397B A17B 28.6 from 1 benchmark.
Widest category
Agentic work, 31.7 points apart on average over 1 shared benchmark.
Widest single gap
FrontierMath, tiers 1 to 3: GLM 5.3 68.8% against Qwen3.5 397B A17B 31.2%, 37.5 points apart on a 0 to 100 scale.
Output price
GLM 5.3 $1.40 against Qwen3.5 397B A17B $2.34 per million output tokens, 1.7x apart.
Context window
GLM 5.3 takes 1.05M tokens, Qwen3.5 397B A17B 262K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGLM 5.3BarsQwen3.5 397B A17BTies
Coding70
Intelligence20
Agentic work10
Reasoning70
Math40
Knowledge40
Web design70
General200
All shared boards520
CodingLike-for-likeFour of these boards feed the Vibe Coding Index70
Coding benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
LiveBench coding82.8%Not measuredOnly GLM 5.3 measured
LMArena WebDev16211400GLM 5.3
AA Coding Index74.8Not measuredOnly GLM 5.3 measured
LiveCodeBench (Vals AI)80.5%Not measuredOnly GLM 5.3 measured
SWE-bench Verified (Vals AI)95.4%Not measuredOnly GLM 5.3 measured
Terminal-Bench 4 (Vals AI)38.9%Not measuredOnly GLM 5.3 measured
Vibe Code Bench (Vals AI)78.1%Not measuredOnly GLM 5.3 measured
Code migration (Vals AI)44.2%Not measuredOnly GLM 5.3 measured
ProgramBench (Vals AI)1.5%Not measuredOnly GLM 5.3 measured
IOI (Vals AI)68.4%Not measuredOnly GLM 5.3 measured
Terminal-Bench 2.1 (Vals AI)71.5%Not measuredOnly GLM 5.3 measured
Vibe Code Bench 1-100 (Vals AI)20.0%Not measuredOnly GLM 5.3 measured
SkillsBench (Vals AI)47.5%Not measuredOnly GLM 5.3 measured
CyberBench (Vals AI)72.1%Not measuredOnly GLM 5.3 measured
SciCode59.0%44.8%GLM 5.3
Terminal-Bench Hard (Artificial Analysis run)Not measured40.9%Only Qwen3.5 397B A17B measured
FrontierCode40.1%Not measuredOnly GLM 5.3 measured
FrontierSWE30.2%Not measuredOnly GLM 5.3 measured
WeirdML75.4%Not measuredOnly GLM 5.3 measured
ALE-Bench1317.4621.9GLM 5.3
LMArena Text: coding14971464GLM 5.3
LMArena Text: software and IT14931463GLM 5.3
LMArena WebDev: React16411388GLM 5.3
LMArena WebDev: HTML15951410GLM 5.3
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
AA Intelligence Index44.821.4GLM 5.3
Epoch Capabilities Index156.0146.8GLM 5.3
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
Agentic work benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
AA Agentic Index53.1Not measuredOnly GLM 5.3 measured
Design Arena: agents building mobile apps1194Not measuredOnly GLM 5.3 measured
Design Arena: HTML slides1189Not measuredOnly GLM 5.3 measured
Design Arena: PowerPoint slides1255Not measuredOnly GLM 5.3 measured
LiveBench research73.2%Not measuredOnly GLM 5.3 measured
LiveBench browser0.0%Not measuredOnly GLM 5.3 measured
LiveBench long agentic35.9%Not measuredOnly GLM 5.3 measured
SREBench (Vals AI)9.2%Not measuredOnly GLM 5.3 measured
Finance Agent (Vals AI)55.8%Not measuredOnly GLM 5.3 measured
Tax Agent Bench (Vals AI)39.7%Not measuredOnly GLM 5.3 measured
Legal Research Bench (Vals AI)49.0%Not measuredOnly GLM 5.3 measured
Harvey's Legal Agent Benchmark (Vals AI)14.6%Not measuredOnly GLM 5.3 measured
Excel Modeling Benchmark (Vals AI)56.3%Not measuredOnly GLM 5.3 measured
Terminal-Bench-Science (Vals AI)5.7%Not measuredOnly GLM 5.3 measured
τ²-Bench TelecomNot measured95.6%Only Qwen3.5 397B A17B measured
APEX-Agents56.6%24.9%GLM 5.3
Vending-Bench 2$8,164Not measuredOnly GLM 5.3 measured
Neither model has a result on the other 22 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index70
Reasoning benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
GPQA Diamond88.1%Not measuredOnly GLM 5.3 measured
LiveBench reasoning85.8%Not measuredOnly GLM 5.3 measured
MysteryMechanism (Vals AI)23.0%Not measuredOnly GLM 5.3 measured
GPQA Diamond (Artificial Analysis run)91.7%89.3%GLM 5.3
Humanity's Last Exam (Artificial Analysis run)42.3%29.0%GLM 5.3
GPQA Diamond (Epoch AI run)90.9%86.4%GLM 5.3
Chess puzzles37.0%Not measuredOnly GLM 5.3 measured
Mystery game puzzles33.0%18.0%GLM 5.3
LMArena Text: hard prompts14901448GLM 5.3
LMArena Text: expert15141462GLM 5.3
LMArena Text: hard prompts in English14961450GLM 5.3
LMArena Vision: diagramsNot measured1277Only Qwen3.5 397B A17B measured
LMArena Vision: homeworkNot measured1292Only Qwen3.5 397B A17B measured
Neither model has a result on the other 8 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index40
Math benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
OTIS Mock AIME91.1%88.9%GLM 5.3
FrontierMath, tiers 1 to 368.8%31.2%GLM 5.3
FrontierMath, tier 429.3%Not measuredOnly GLM 5.3 measured
ProofBench49.0%Not measuredOnly GLM 5.3 measured
LMArena Text: math14951453GLM 5.3
LMArena Text: mathematical work14971454GLM 5.3
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index40
Knowledge benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
MMLU Pro86.8%Not measuredOnly GLM 5.3 measured
LegalBench (Vals AI)84.8%Not measuredOnly GLM 5.3 measured
TaxEval (Vals AI)72.4%Not measuredOnly GLM 5.3 measured
MedCode (Vals AI)42.9%Not measuredOnly GLM 5.3 measured
SimpleQA Verified41.0%Not measuredOnly GLM 5.3 measured
LMArena Text: science14981457GLM 5.3
LMArena Text: medicine and healthcare14811455GLM 5.3
LMArena Text: legal and government14731441GLM 5.3
LMArena Text: business and finance14591435GLM 5.3
Neither model has a result on the other 8 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index70
Web design benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
Design Arena: websites12951191GLM 5.3
Design Arena: UI components13141163GLM 5.3
Design Arena: game dev13111141GLM 5.3
Design Arena: data viz12601175GLM 5.3
Design Arena: 3D13321161GLM 5.3
Design Arena: all code categories13041180GLM 5.3
Design Arena: SVG12921139GLM 5.3
Design Arena: ASCII art1213Not measuredOnly GLM 5.3 measured
LiveBench design77.3%Not measuredOnly GLM 5.3 measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index200
General benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
Vals Index59.7%Not measuredOnly GLM 5.3 measured
LiveBench data analysis70.2%Not measuredOnly GLM 5.3 measured
Public Benefits Bench (Vals AI)68.5%Not measuredOnly GLM 5.3 measured
MedScribe (Vals AI)88.8%Not measuredOnly GLM 5.3 measured
IFBenchNot measured78.8%Only Qwen3.5 397B A17B measured
Long context reasoning (AA-LCR)79.7%77.3%GLM 5.3
DTBench87.7%87.5%GLM 5.3
LMCA55.5%37.9%GLM 5.3
LMArena Text14741437GLM 5.3
LMArena Text: instruction following14791424GLM 5.3
LMArena Text: creative writing14581402GLM 5.3
LMArena VisionNot measured1262Only Qwen3.5 397B A17B measured
LMArena Text: English14821443GLM 5.3
LMArena Text: non-English14591430GLM 5.3
LMArena Text: Chinese15321498GLM 5.3
LMArena Text: German14881445GLM 5.3
LMArena Text: Japanese14601429GLM 5.3
LMArena Text: Korean14601384GLM 5.3
LMArena Text: French15031462GLM 5.3
LMArena Text: Spanish14621441GLM 5.3
LMArena Text: Russian14691430GLM 5.3
LMArena Text: Polish14661447GLM 5.3
LMArena Text: multi-turn14721446GLM 5.3
LMArena Text: longer queries14831441GLM 5.3
LMArena Text: writing and language14691415GLM 5.3
LMArena Text: entertainment and media14471399GLM 5.3
LMArena Vision: EnglishNot measured1268Only Qwen3.5 397B A17B measured
LMArena Vision: ChineseNot measured1314Only Qwen3.5 397B A17B measured
LMArena Vision: reading textNot measured1271Only Qwen3.5 397B A17B measured
LMArena Vision: humorNot measured1262Only Qwen3.5 397B A17B measured
LMArena Vision: captioningNot measured1230Only Qwen3.5 397B A17B measured
LMArena Vision: recognizing thingsNot measured1265Only Qwen3.5 397B A17B measured
LMArena Vision: creative writingNot measured1260Only Qwen3.5 397B A17B measured
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GLM 5.3 and Qwen3.5 397B A17B
BenchmarkGLM 5.3Qwen3.5 397B A17BLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GLM 5.3$0.0008Fits in one request
Qwen3.5 397B A17B$0.0016Fits in one request

GLM 5.3 costs less

Repository review

50K input + 3K output tokens

GLM 5.3$0.0081Fits in one request
Qwen3.5 397B A17B$0.027Fits in one request

GLM 5.3 costs less

Agent loop

200K input + 10K output tokens

GLM 5.3$0.030Fits in one request
Qwen3.5 397B A17B$0.101Fits in one request

GLM 5.3 costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GLM 5.3 and Qwen3.5 397B A17B
SpecificationZ.ai GLM 5.3Qwen3.5 397B A17B
MakerZ.aiQwen
WeightsOpen weightsOpen weights
ReleasedAugust 18, 2026February 16, 2026
Context window1.05M262K
Input $/Mtok$0.079$0.39
Output $/Mtok$1.40$2.34
Coding index73.128.6
Independent measurements behind it21
Cost per solved task$0.36Not computable
SourceOpenRouter listingOpenRouter listing

Questions

Is GLM 5.3 better than Qwen3.5 397B A17B for coding?

Pick GLM 5.3 for coding: it scores 73.1 on the Vibe Coding Index against 28.6 for Qwen3.5 397B A17B. Qwen3.5 397B A17B is neither cheaper nor larger in context, and leads no benchmark both models were measured on.

Which is cheaper, GLM 5.3 or Qwen3.5 397B A17B?

GLM 5.3 costs $0.079 per million input tokens and $1.40 per million output tokens; Qwen3.5 397B A17B costs $0.39 and $2.34. Cost per solved task is known for GLM 5.3 only.

Which has the bigger context window, GLM 5.3 or Qwen3.5 397B A17B?

GLM 5.3, with 1.05M tokens against 262K.

How many benchmarks is the GLM 5.3 vs Qwen3.5 397B A17B comparison based on?

52 benchmarks have a result for both models as of October 21, 2026. GLM 5.3's coding index rests on 2 independent benchmarks and Qwen3.5 397B A17B's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next