Enter opens the highlighted result. Esc closes.

GLM 5.3 Flash vs GPT-5.4 Mini

Updated October 21, 2026. 73 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 Flash Replace GPT-5.4 Mini
GLM 5.3 FlashZ.ai
Coding
71.0
Overall
41.8
Output $/M
$0.50
65
GPT-5.4 MiniOpenAI
Coding
28.2
Overall
24.1
Output $/M
$4.50
6
vs boards won of 73 shared, 2 tied

Decision reading

Pick GLM 5.3 Flash for coding: it scores 71.0 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini.

Pick GPT-5.4 Mini when what matters is LiveCodeBench (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Model AGLM 5.3 FlashZ.ai71.0/1002 independent benchmarks$0.02 per solved taskModel BGPT-5.4 MiniOpenAI28.2/1001 independent benchmark$0.70 per solved task
Shared results
73
GLM 5.3 Flash only
29
GPT-5.4 Mini only
15
Like-for-like categories
7 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 Flash 71.0 from 2 benchmarks, GPT-5.4 Mini 28.2 from 1 benchmark.
Widest category
Agentic work, 22.3 points apart on average over 5 shared benchmarks.
Widest single gap
ARC-AGI-2: GLM 5.3 Flash 65.8% against GPT-5.4 Mini 18.9%, 46.9 points apart on a 0 to 100 scale.
Output price
GLM 5.3 Flash $0.50 against GPT-5.4 Mini $4.50 per million output tokens, 9.0x apart.
Context window
GLM 5.3 Flash takes 1.05M tokens, GPT-5.4 Mini 400K.

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 5.3 FlashBarsGPT-5.4 MiniTies
Coding1141 tie
Intelligence20
Agentic work50
Reasoning122
Math501 tie
Knowledge60
General240
All shared boards6562 ties
CodingLike-for-likeFour of these boards feed the Vibe Coding Index114
Coding benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
LiveBench coding80.6%Not measuredOnly GLM 5.3 Flash measured
LMArena WebDev16071397GLM 5.3 Flash
AA Coding Index71.556.1GLM 5.3 Flash
LiveCodeBench (Vals AI)80.5%81.5%GPT-5.4 Mini
SWE-bench Verified (Vals AI)92.0%73.0%GLM 5.3 Flash
Terminal-Bench 4 (Vals AI)37.4%2.5%GLM 5.3 Flash
Vibe Code Bench (Vals AI)30.8%48.0%GPT-5.4 Mini
Code migration (Vals AI)20.5%12.9%GLM 5.3 Flash
ProgramBench (Vals AI)0.0%0.0%Tie
IOI (Vals AI)52.5%Not measuredOnly GLM 5.3 Flash measured
Terminal-Bench 2.1 (Vals AI)62.9%54.7%GLM 5.3 Flash
Vibe Code Bench 1-100 (Vals AI)16.0%Not measuredOnly GLM 5.3 Flash measured
SkillsBench (Vals AI)40.2%Not measuredOnly GLM 5.3 Flash measured
CyberBench (Vals AI)67.9%Not measuredOnly GLM 5.3 Flash measured
SciCode51.6%52.1%GPT-5.4 Mini
Terminal-Bench Hard (Artificial Analysis run)Not measured52.3%Only GPT-5.4 Mini measured
FrontierCode31.8%27.0%GLM 5.3 Flash
FrontierSWE18.1%Not measuredOnly GLM 5.3 Flash measured
WeirdMLNot measured60.3%Only GPT-5.4 Mini measured
ALE-Bench303.61188.6GPT-5.4 Mini
LMArena Text: coding15071438GLM 5.3 Flash
LMArena Text: software and IT15001436GLM 5.3 Flash
LMArena WebDev: React16241386GLM 5.3 Flash
LMArena WebDev: HTML15771419GLM 5.3 Flash
LMArena WebDev: from an image1578Not measuredOnly GLM 5.3 Flash measured
Neither model has a result on the other 13 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
AA Intelligence Index41.824.1GLM 5.3 Flash
Epoch Capabilities Index153.1148.7GLM 5.3 Flash
Agentic workLike-for-likePublic boards, shown as published, kept out of the index50
Agentic work benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
AA Agentic Index50.917.9GLM 5.3 Flash
LiveBench research70.1%Not measuredOnly GLM 5.3 Flash measured
LiveBench browser84.0%Not measuredOnly GLM 5.3 Flash measured
LiveBench long agentic28.4%Not measuredOnly GLM 5.3 Flash measured
SREBench (Vals AI)0.8%Not measuredOnly GLM 5.3 Flash measured
Finance Agent (Vals AI)57.9%45.4%GLM 5.3 Flash
Tax Agent Bench (Vals AI)31.6%8.8%GLM 5.3 Flash
Legal Research Bench (Vals AI)45.2%12.5%GLM 5.3 Flash
Harvey's Legal Agent Benchmark (Vals AI)10.4%Not measuredOnly GLM 5.3 Flash measured
Excel Modeling Benchmark (Vals AI)55.9%45.4%GLM 5.3 Flash
Terminal-Bench-Science (Vals AI)2.9%Not measuredOnly GLM 5.3 Flash measured
τ²-Bench TelecomNot measured83.3%Only GPT-5.4 Mini measured
APEX-Agents52.8%Not measuredOnly GLM 5.3 Flash measured
DeepResearch BenchNot measured36.3%Only GPT-5.4 Mini measured
Neither model has a result on the other 25 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index122
Reasoning benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
GPQA Diamond86.4%83.1%GLM 5.3 Flash
LiveBench reasoning77.6%Not measuredOnly GLM 5.3 Flash measured
MMMU Pro (Vals AI)86.0%79.3%GLM 5.3 Flash
GPQA Diamond (Artificial Analysis run)91.2%87.5%GLM 5.3 Flash
Humanity's Last Exam (Artificial Analysis run)39.9%28.1%GLM 5.3 Flash
GPQA Diamond (Epoch AI run)90.2%86.9%GLM 5.3 Flash
ARC-AGI91.0%63.7%GLM 5.3 Flash
ARC-AGI-265.8%18.9%GLM 5.3 Flash
Chess puzzles21.0%23.0%GPT-5.4 Mini
Mystery game puzzles8.0%11.0%GPT-5.4 Mini
LMArena Text: hard prompts14901424GLM 5.3 Flash
LMArena Text: expert15111435GLM 5.3 Flash
LMArena Text: hard prompts in English14881420GLM 5.3 Flash
LMArena Vision: diagrams13171266GLM 5.3 Flash
LMArena Vision: homework13181285GLM 5.3 Flash
Neither model has a result on the other 6 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index50
Math benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
AIMENot measured95.6%Only GPT-5.4 Mini measured
OTIS Mock AIME93.9%88.9%GLM 5.3 Flash
FrontierMath, tiers 1 to 355.8%51.2%GLM 5.3 Flash
FrontierMath, tier 417.1%9.8%GLM 5.3 Flash
ProofBench21.0%21.0%Tie
LMArena Text: math15031418GLM 5.3 Flash
LMArena Text: mathematical work14971429GLM 5.3 Flash
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index60
Knowledge benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
MMLU Pro86.1%84.5%GLM 5.3 Flash
LegalBench (Vals AI)83.9%Not measuredOnly GLM 5.3 Flash measured
CaseLaw (Vals AI)Not measured51.7%Only GPT-5.4 Mini measured
CorpFin (Vals AI)Not measured60.9%Only GPT-5.4 Mini measured
TaxEval (Vals AI)75.6%71.2%GLM 5.3 Flash
MortgageTax (Vals AI)Not measured63.5%Only GPT-5.4 Mini measured
SimpleQA VerifiedNot measured29.4%Only GPT-5.4 Mini measured
LMArena Text: science14841417GLM 5.3 Flash
LMArena Text: medicine and healthcare14751395GLM 5.3 Flash
LMArena Text: legal and government14731414GLM 5.3 Flash
LMArena Text: business and finance14601418GLM 5.3 Flash
Neither model has a result on the other 6 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
Design Arena: websites1272Not measuredOnly GLM 5.3 Flash measured
Design Arena: UI components1316Not measuredOnly GLM 5.3 Flash measured
Design Arena: game dev1279Not measuredOnly GLM 5.3 Flash measured
Design Arena: data viz1262Not measuredOnly GLM 5.3 Flash measured
Design Arena: 3D1316Not measuredOnly GLM 5.3 Flash measured
Design Arena: all code categories1282Not measuredOnly GLM 5.3 Flash measured
Design Arena: SVG1284Not measuredOnly GLM 5.3 Flash measured
Design Arena: ASCII art1269Not measuredOnly GLM 5.3 Flash measured
LiveBench design77.2%Not measuredOnly GLM 5.3 Flash measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index240
General benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
Vals Index50.9%Not measuredOnly GLM 5.3 Flash measured
LiveBench data analysis76.4%Not measuredOnly GLM 5.3 Flash measured
SAGE (Vals AI)Not measured50.8%Only GPT-5.4 Mini measured
MedScribe (Vals AI)88.9%Not measuredOnly GLM 5.3 Flash measured
Vals Multimodal IndexNot measured54.2%Only GPT-5.4 Mini measured
IFBenchNot measured73.3%Only GPT-5.4 Mini measured
Long context reasoning (AA-LCR)80.0%77.0%GLM 5.3 Flash
DTBenchNot measured80.0%Only GPT-5.4 Mini measured
LMCANot measured40.8%Only GPT-5.4 Mini measured
Surface Evolver Bench52.5%Not measuredOnly GLM 5.3 Flash measured
LMArena Text14711412GLM 5.3 Flash
LMArena Text: instruction following14781405GLM 5.3 Flash
LMArena Text: creative writing14411370GLM 5.3 Flash
LMArena Vision12941245GLM 5.3 Flash
LMArena Text: English14741412GLM 5.3 Flash
LMArena Text: non-English14621405GLM 5.3 Flash
LMArena Text: Chinese15311446GLM 5.3 Flash
LMArena Text: German14621408GLM 5.3 Flash
LMArena Text: Japanese14441375GLM 5.3 Flash
LMArena Text: Korean14361370GLM 5.3 Flash
LMArena Text: French14911441GLM 5.3 Flash
LMArena Text: Spanish14751405GLM 5.3 Flash
LMArena Text: Russian14651417GLM 5.3 Flash
LMArena Text: Polish14501415GLM 5.3 Flash
LMArena Text: multi-turn14691429GLM 5.3 Flash
LMArena Text: longer queries14811408GLM 5.3 Flash
LMArena Text: writing and language14551393GLM 5.3 Flash
LMArena Text: entertainment and media14401373GLM 5.3 Flash
LMArena Vision: English12951244GLM 5.3 Flash
LMArena Vision: Chinese13631286GLM 5.3 Flash
LMArena Vision: reading text12991253GLM 5.3 Flash
LMArena Vision: humor12791235GLM 5.3 Flash
LMArena Vision: recognizing thingsNot measured1218Only GPT-5.4 Mini measured
LMArena Vision: creative writing12701222GLM 5.3 Flash
Neither model has a result on the other 12 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GLM 5.3 Flash and GPT-5.4 Mini
BenchmarkGLM 5.3 FlashGPT-5.4 MiniLeader
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 Flash$0.0004Fits in one request
GPT-5.4 Mini$0.0030Fits in one request

GLM 5.3 Flash costs less

Repository review

50K input + 3K output tokens

GLM 5.3 Flash$0.0090Fits in one request
GPT-5.4 Mini$0.051Fits in one request

GLM 5.3 Flash costs less

Agent loop

200K input + 10K output tokens

GLM 5.3 Flash$0.035Fits in one request
GPT-5.4 Mini$0.195Fits in one request

GLM 5.3 Flash 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 Flash and GPT-5.4 Mini
SpecificationZ.ai GLM 5.3 FlashOpenAI GPT-5.4 Mini
MakerZ.aiOpenAI
WeightsOpen weightsNo public weights listed
ReleasedAugust 26, 2026March 17, 2026
Context window1.05M400K
Input $/Mtok$0.15$0.75
Output $/Mtok$0.50$4.50
Coding index71.028.2
Independent measurements behind it21
Cost per solved task$0.02$0.70
SourceOpenRouter listingOpenRouter listing

Questions

Is GLM 5.3 Flash better than GPT-5.4 Mini for coding?

Pick GLM 5.3 Flash for coding: it scores 71.0 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini. Pick GPT-5.4 Mini when what matters is LiveCodeBench (Vals AI) or Vibe Code Bench (Vals AI), where it leads.

Which is cheaper, GLM 5.3 Flash or GPT-5.4 Mini?

GLM 5.3 Flash costs $0.15 per million input tokens and $0.50 per million output tokens; GPT-5.4 Mini costs $0.75 and $4.50. Per solved SWE-bench Verified task, GLM 5.3 Flash costs $0.02 and GPT-5.4 Mini $0.70.

Which has the bigger context window, GLM 5.3 Flash or GPT-5.4 Mini?

GLM 5.3 Flash, with 1.05M tokens against 400K.

How many benchmarks is the GLM 5.3 Flash vs GPT-5.4 Mini comparison based on?

73 benchmarks have a result for both models as of October 21, 2026. GLM 5.3 Flash's coding index rests on 2 independent benchmarks and GPT-5.4 Mini's on 1. Where only one model has a result, the page says so instead of naming a winner.

Compare next