LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-5.3-Codex vs Kimi K2.5

Updated October 21, 2026. 32 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-5.3-Codex Replace Kimi K2.5
GPT-5.3-CodexOpenAI
Coding
29.5
Overall
32.5
Output $/M
$14.00
15
Kimi K2.5Moonshot AI
Coding
58.5
Overall
23.5
Output $/M
$2.50
17
vs boards won of 32 shared

Decision reading

Pick Kimi K2.5 for coding: it scores 58.5 on the Vibe Coding Index against 29.5 for GPT-5.3-Codex.

Pick GPT-5.3-Codex when the context window is larger, 400K against 262K tokens, or when what matters is LiveCodeBench (Vals AI) or SWE-bench Verified (Vals AI), where it leads.

Model AGPT-5.3-CodexOpenAI29.5/1001 independent benchmark$0.59 per solved taskModel BKimi K2.5Moonshot AI58.5/1002 independent benchmarks$0.24 per solved task
Shared results
32
GPT-5.3-Codex only
2
Kimi K2.5 only
66
Like-for-like categories
6 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
GPT-5.3-Codex 29.5 from 1 benchmark, Kimi K2.5 58.5 from 2 benchmarks.
Widest category
Coding, 16.5 points apart on average over 11 shared benchmarks.
Widest single gap
Vibe Code Bench (Vals AI): GPT-5.3-Codex 61.8% against Kimi K2.5 17.5%, 44.2 points apart on a 0 to 100 scale.
Output price
GPT-5.3-Codex $14.00 against Kimi K2.5 $2.50 per million output tokens, 5.6x apart.
Context window
GPT-5.3-Codex takes 400K tokens, Kimi K2.5 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
CategoryGPT-5.3-CodexBarsKimi K2.5Ties
Coding83
Intelligence20
Agentic work16
Reasoning20
Web design08
General20
All shared boards1517
CodingLike-for-likeFour of these boards feed the Vibe Coding Index83
Coding benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
SWE-bench VerifiedNot measured70.8%Only Kimi K2.5 measured
SWE-bench MultilingualNot measured67.3%Only Kimi K2.5 measured
LMArena WebDev14071437Kimi K2.5
AA Coding IndexNot measured46.8Only Kimi K2.5 measured
LiveCodeBench (Vals AI)87.3%83.9%GPT-5.3-Codex
SWE-bench Verified (Vals AI)78.0%70.0%GPT-5.3-Codex
Vibe Code Bench (Vals AI)61.8%17.5%GPT-5.3-Codex
Code migration (Vals AI)Not measured7.0%Only Kimi K2.5 measured
IOI (Vals AI)53.8%Not measuredOnly GPT-5.3-Codex measured
Terminal-Bench 2.1 (Vals AI)Not measured42.0%Only Kimi K2.5 measured
Terminal-Bench Hard (Artificial Analysis run)53.0%34.8%GPT-5.3-Codex
Terminal-Bench78.4%43.2%GPT-5.3-Codex
WeirdML79.3%45.6%GPT-5.3-Codex
ALE-Bench1655.2821.6GPT-5.3-Codex
LMArena Text: codingNot measured1483Only Kimi K2.5 measured
LMArena Text: software and ITNot measured1471Only Kimi K2.5 measured
LMArena WebDev: React13951429Kimi K2.5
LMArena WebDev: HTML14221426Kimi K2.5
LMArena WebDev: from an image14431439GPT-5.3-Codex
Neither model has a result on the other 19 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
AA Intelligence Index32.523.5GPT-5.3-Codex
Epoch Capabilities Index157.1148.1GPT-5.3-Codex
Agentic workLike-for-likePublic boards, shown as published, kept out of the index16
Agentic work benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
Design Arena: agents building web apps10331136Kimi K2.5
Design Arena: agents building full-stack apps9831114Kimi K2.5
Design Arena: agents building mobile apps10541147Kimi K2.5
Design Arena: agents building Android apps10841124Kimi K2.5
Design Arena: agents building Godot games11241219Kimi K2.5
Finance Agent (Vals AI)Not measured35.8%Only Kimi K2.5 measured
Legal Research Bench (Vals AI)Not measured15.9%Only Kimi K2.5 measured
Excel Modeling Benchmark (Vals AI)Not measured28.5%Only Kimi K2.5 measured
τ²-Bench Telecom86.0%95.9%Kimi K2.5
METR time horizons74.5%Not measuredOnly GPT-5.3-Codex measured
OSWorldNot measured63.3%Only Kimi K2.5 measured
Vending-Bench 2$5,940$1,198GPT-5.3-Codex
Neither model has a result on the other 27 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
Reasoning benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
GPQA DiamondNot measured84.1%Only Kimi K2.5 measured
MMMU Pro (Vals AI)Not measured84.3%Only Kimi K2.5 measured
GPQA Diamond (Artificial Analysis run)91.5%87.9%GPT-5.3-Codex
Humanity's Last Exam (Artificial Analysis run)42.5%30.7%GPT-5.3-Codex
ARC-AGINot measured65.3%Only Kimi K2.5 measured
ARC-AGI-2Not measured11.8%Only Kimi K2.5 measured
SimpleBenchNot measured46.8%Only Kimi K2.5 measured
LMArena Text: hard promptsNot measured1453Only Kimi K2.5 measured
LMArena Text: expertNot measured1466Only Kimi K2.5 measured
LMArena Text: hard prompts in EnglishNot measured1460Only Kimi K2.5 measured
LMArena Vision: diagramsNot measured1282Only Kimi K2.5 measured
LMArena Vision: homeworkNot measured1297Only Kimi K2.5 measured
Neither model has a result on the other 9 boards here
MathDirectional onlyPublic boards, shown as published, kept out of the index3 boards, none shared
Math benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
AIMENot measured95.6%Only Kimi K2.5 measured
LMArena Text: mathNot measured1470Only Kimi K2.5 measured
LMArena Text: mathematical workNot measured1473Only Kimi K2.5 measured
Neither model has a result on the other 10 boards here
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
Knowledge benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
MMLU ProNot measured85.9%Only Kimi K2.5 measured
CaseLaw (Vals AI)Not measured58.7%Only Kimi K2.5 measured
CorpFin (Vals AI)Not measured68.3%Only Kimi K2.5 measured
TaxEval (Vals AI)Not measured74.2%Only Kimi K2.5 measured
MortgageTax (Vals AI)Not measured66.5%Only Kimi K2.5 measured
MedQA (Vals AI)Not measured94.4%Only Kimi K2.5 measured
MedCode (Vals AI)Not measured39.3%Only Kimi K2.5 measured
SimpleQA VerifiedNot measured34.3%Only Kimi K2.5 measured
LMArena Text: scienceNot measured1459Only Kimi K2.5 measured
LMArena Text: medicine and healthcareNot measured1448Only Kimi K2.5 measured
LMArena Text: legal and governmentNot measured1441Only Kimi K2.5 measured
LMArena Text: business and financeNot measured1427Only Kimi K2.5 measured
Neither model has a result on the other 5 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index08
Web design benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
Design Arena: websites11631249Kimi K2.5
Design Arena: UI components11371236Kimi K2.5
Design Arena: game dev11621207Kimi K2.5
Design Arena: data viz11641219Kimi K2.5
Design Arena: 3D10081205Kimi K2.5
Design Arena: all code categories11451236Kimi K2.5
Design Arena: SVG11291149Kimi K2.5
Design Arena: ASCII art11511165Kimi K2.5
Neither model has a result on the other 1 board here
GeneralLike-for-likePublic boards, shown as published, kept out of the index20
General benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
SAGE (Vals AI)Not measured49.9%Only Kimi K2.5 measured
MedScribe (Vals AI)Not measured76.4%Only Kimi K2.5 measured
IFBench75.4%70.2%GPT-5.3-Codex
Long context reasoning (AA-LCR)83.3%78.0%GPT-5.3-Codex
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena TextNot measured1445Only Kimi K2.5 measured
LMArena Text: instruction followingNot measured1431Only Kimi K2.5 measured
LMArena Text: creative writingNot measured1422Only Kimi K2.5 measured
LMArena VisionNot measured1269Only Kimi K2.5 measured
LMArena Text: EnglishNot measured1453Only Kimi K2.5 measured
LMArena Text: non-EnglishNot measured1433Only Kimi K2.5 measured
LMArena Text: ChineseNot measured1494Only Kimi K2.5 measured
LMArena Text: GermanNot measured1441Only Kimi K2.5 measured
LMArena Text: JapaneseNot measured1423Only Kimi K2.5 measured
LMArena Text: KoreanNot measured1412Only Kimi K2.5 measured
LMArena Text: FrenchNot measured1454Only Kimi K2.5 measured
LMArena Text: SpanishNot measured1450Only Kimi K2.5 measured
LMArena Text: RussianNot measured1435Only Kimi K2.5 measured
LMArena Text: PolishNot measured1445Only Kimi K2.5 measured
LMArena Text: multi-turnNot measured1444Only Kimi K2.5 measured
LMArena Text: longer queriesNot measured1445Only Kimi K2.5 measured
LMArena Text: writing and languageNot measured1425Only Kimi K2.5 measured
LMArena Text: entertainment and mediaNot measured1417Only Kimi K2.5 measured
LMArena Vision: EnglishNot measured1274Only Kimi K2.5 measured
LMArena Vision: ChineseNot measured1328Only Kimi K2.5 measured
LMArena Vision: reading textNot measured1276Only Kimi K2.5 measured
LMArena Vision: humorNot measured1284Only Kimi K2.5 measured
LMArena Vision: captioningNot measured1257Only Kimi K2.5 measured
LMArena Vision: recognizing thingsNot measured1285Only Kimi K2.5 measured
LMArena Vision: creative writingNot measured1269Only Kimi K2.5 measured
Neither model has a result on the other 14 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.3-Codex and Kimi K2.5
BenchmarkGPT-5.3-CodexKimi K2.5Leader
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

GPT-5.3-Codex$0.0088Fits in one request
Kimi K2.5$0.0018Fits in one request

Kimi K2.5 costs less

Repository review

50K input + 3K output tokens

GPT-5.3-Codex$0.130Fits in one request
Kimi K2.5$0.033Fits in one request

Kimi K2.5 costs less

Agent loop

200K input + 10K output tokens

GPT-5.3-Codex$0.490Fits in one request
Kimi K2.5$0.125Fits in one request

Kimi K2.5 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 GPT-5.3-Codex and Kimi K2.5
SpecificationOpenAI GPT-5.3-CodexMoonshotAI Kimi K2.5
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedFebruary 24, 2026January 27, 2026
Context window400K262K
Input $/Mtok$1.75$0.50
Output $/Mtok$14.00$2.50
Coding index29.558.5
Independent measurements behind it12
Cost per solved task$0.59$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.3-Codex better than Kimi K2.5 for coding?

Pick Kimi K2.5 for coding: it scores 58.5 on the Vibe Coding Index against 29.5 for GPT-5.3-Codex. Pick GPT-5.3-Codex when the context window is larger, 400K against 262K tokens, or when what matters is LiveCodeBench (Vals AI) or SWE-bench Verified (Vals AI), where it leads.

Which is cheaper, GPT-5.3-Codex or Kimi K2.5?

GPT-5.3-Codex costs $1.75 per million input tokens and $14.00 per million output tokens; Kimi K2.5 costs $0.50 and $2.50. Per solved SWE-bench Verified task, GPT-5.3-Codex costs $0.59 and Kimi K2.5 $0.24.

Which has the bigger context window, GPT-5.3-Codex or Kimi K2.5?

GPT-5.3-Codex, with 400K tokens against 262K.

How many benchmarks is the GPT-5.3-Codex vs Kimi K2.5 comparison based on?

32 benchmarks have a result for both models as of October 21, 2026. GPT-5.3-Codex's coding index rests on 1 independent benchmark and Kimi K2.5's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next