BenchLeader

GLM 5.3 vs Kimi K3

Full profiles: GLM 5.3 (Zhipu AI) and Kimi K3 (Moonshot AI): every benchmark, price and rank.

Updated · Data from 15 sources

As of 16 Nov 2026, GLM 5.3 and Kimi K3 are neck and neck on quality (58.8 vs 60.7 on the BenchLeader Index, inside the ±6.7 margin), and GLM 5.3 is 2.8× cheaper at $2.15 vs $6.00 per 1M tokens blended. Choose GLM 5.3 for agents & tools (58 vs 54); choose Kimi K3 for fast first answers (0.8 s vs 57.6 s) and knowledge (66 vs 54), maths (72 vs 60), and coding (62 vs 60).

Key facts (as of 16 Nov 2026)

  • GLM 5.3 scores 58.8 on the BenchLeader Index and Kimi K3 scores 60.7 (BenchLeader, 16 Nov 2026).
  • GLM 5.3 costs $2.15 per 1M tokens blended (list price, 16 Nov 2026).
  • Kimi K3 costs $6.00 per 1M tokens blended (list price, 16 Nov 2026).
  • Of the 10 benchmarks both models have, Kimi K3 leads 7 (BenchLeader, 16 Nov 2026).

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use casePickWhyConfidence
CodingKimi K362 vs 60 category score; FrontierCode 44.2% vs 40.1%Medium
Agents & tool useGLM 5.358 vs 54 category score; Terminal-Bench 2.1 (Vals) 71.5% vs 80.9%Medium
MathsKimi K372 vs 60 category score; ProofBench 87.0% vs 49.0%High
KnowledgeKimi K366 vs 54 category score; TaxEval 75.7% vs 72.4%High
Chat quality (human preference)Kimi K369 vs 67 on Arena votesMedium
Lowest costGLM 5.3$2.15 vs $6.00 per 1M blended; agentic coding session $0.102 vs $0.240–
Fastest first answerKimi K30.8 s (Kimi K3) vs 57.6 s (GLM 5.3 (max)) to the first answer token–
High volume on a budgetNeither flagshipA smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs–

Key differences

  • GLM 5.3 (max) and Kimi K3 are neck and neck on quality (58.8 vs 60.7, within the ±6.7 margin).
  • GLM 5.3 (max) is stronger in agents & tools.
  • Kimi K3 is stronger in coding, human preference, knowledge, maths.
  • GLM 5.3 (max) is 2.8× cheaper ($2.15 vs $6.00 per 1M blended).

At a glance

SpecGLM 5.3Kimi K3
DeveloperZhipu AIMoonshot AI
Released18 Aug 202616 Jul 2026
Context window (tokens)1,000,0001,048,576
Compared atGLM 5.3 (max)Kimi K3
Open weightsYesYes
Price in / out / cached input per 1M$1.40 / $4.40 / $0.260$3.00 / $15.00 / $0.300

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

CategoryGLM 5.3 (max)Kimi K3Leader
Reasoning54.7–not comparable
Coding59.761.8Kimi K3
Agents & tools57.653.5GLM 5.3
Maths60.071.5Kimi K3
Knowledge53.566.1Kimi K3
Human preference67.268.6Kimi K3
Long context64.5–not comparable

On one chart

Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.

2550 avg75CodingAgents & toolsMathsKnowledgeHumanpreferenceGLM 5.3, Coding: 59.7GLM 5.3, Agents & tools: 57.6GLM 5.3, Maths: 60.0GLM 5.3, Knowledge: 53.5GLM 5.3, Human preference: 67.2Kimi K3, Coding: 61.8Kimi K3, Agents & tools: 53.5Kimi K3, Maths: 71.5Kimi K3, Knowledge: 66.1Kimi K3, Human preference: 68.6

Not drawn, measured for fewer than two of these models: Reasoning, Instruction following, Long context, Multimodal.

Model, click to add or removeIndex

Where sources disagree

Across the 10 benchmarks both have, GLM 5.3 leads 3, Kimi K3 leads 7. Independent sources split: GLM 5.3 leads on AA Intelligence Index v5.0.1 and Vals Index; Kimi K3 leads on Arena Text.

Overall indexGLM 5.3 (max)Kimi K3Leader
BenchLeader Index58.860.7neck and neck
AA Intelligence Index v5.0.131.617.4†GLM 5.3
Vals Index59.953.0†GLM 5.3
Arena Text14821487†Kimi K3

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.

MeasureGLM 5.3 (max)Kimi K3
Blended price per 1M tokens$2.15$6.00
Input / output per 1M$1.40 / $4.40$3.00 / $15.00
Cached input per 1M$0.260$0.300
Price per task: chat reply$0.0019$0.0057
Price per task: summarize a 30-page report$0.019$0.045
Price per task: code edit$0.015$0.041
Price per task: agentic coding session$0.102$0.240
Price per task: structured extraction$0.0037$0.0090
Agentic coding session, input mostly cached$0.050$0.118
Cost per benchmark task (AA Index)$5.01–
Output tokens per task (AA Index)119k–
Of which reasoning85k–
Time per task (AA Index)23 min–
Output speed88 tok/s76 tok/s
Time to first answer57.6 s0.8 s
Time to a complete answer, last measured 13 Nov 2026144.1 s–

Each model is compared at its best-ranked reasoning-effort setting. GLM 5.3 (max) scores 58.8 and takes 58 s to its first answer token; GLM 5.3 (low) scores 47.3 in 7 s. Kimi K3 scores 60.7 and takes 1 s to its first answer token; Kimi K3 (max) scores 60.2 in 55 s, Kimi K3 (low) scores 53.6 in 8 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

176 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

MetricPublisherGLM 5.3 (max)Kimi K3
BenchLeader IndexBenchLeader58.860.7
Agents & tools scoreBenchLeader57.653.5
Coding scoreBenchLeader59.761.8
Human preference scoreBenchLeader67.268.6
Knowledge scoreBenchLeader53.566.1
Long context scoreBenchLeader64.5–
Maths scoreBenchLeader60.071.5
Reasoning scoreBenchLeader54.7–
Blended price $/MBenchLeader$2.15$6.00
Output speedBenchLeader88 tok/s76 tok/s
Time to first answerBenchLeader57.6 s0.8 s
Context windowBenchLeader1,000,000 tokens1,048,576 tokens
Reasoning
GPQA DiamondEpoch AI Benchmarking Hub90.9%93.1%†
SimpleBenchSimpleBench66.2%†60.7%†
Arena Hard PromptsArena (LMArena)15071516†
LiveBench 2.0 ReasoningLiveBench–90.7%†
LiveBench 1.0 ReasoningLiveBench85.8%†90.7%
GPQA Diamond (AA)Artificial Analysis91.7%93.5%†
Humanity's Last Exam (AA)Artificial Analysis42.3%46.9%†
HLE-Diamond (AA)Artificial Analysis17.9%26.9%†
GPQA Diamond (Vals)Vals AI88.1%92.9%
ARC-AGI-1ARC Prize–94.5%†
ARC-AGI-2ARC Prize–60.4%†
CritPtArtificial Analysis19.1%23.4%†
CritPt v2 hard subset (AA)Artificial Analysis33.5%40.5%†
MysteryMechanismVals AI23.0%–
HLE-Diamond (Scale)Scale AI SEAL16.4%†22.2%
Arena ExpertArena (LMArena)15221528†
Chess PuzzlesEpoch AI Benchmarking Hub37.0%36.0%†
Mystery Game PuzzlesEpoch AI Benchmarking Hub33.0%26.0%†
LMCAEpoch AI Benchmarking Hub55.5%†52.7%†
DTBenchEpoch AI Benchmarking Hub87.7%†91.2%†
ForecastBenchForecastBench–64.2%†
BTF-3Epoch AI Benchmarking Hub14.9%†–
Coding
SciCodeSciCode59.0%58.7%
WeirdMLWeirdML75.4%82.6%†
FrontierCodeCognition40.1%44.2%
Show all 176 benchmarks
MetricPublisherGLM 5.3 (max)Kimi K3
Coding
Arena CodingArena (LMArena)15221538†
Arena WebDevArena (LMArena)16231658†
LiveBench 2.0 CodingLiveBench–53.1%†
LiveBench 1.0 CodingLiveBench79.0%†81.5%
SciCode (AA)Artificial Analysis59.0%59.5%†
LiveCodeBenchVals AI80.5%87.2%†
IOIVals AI68.4%48.9%†
SWE-bench (Vals)Vals AI95.4%93.4%
Code MigrationVals AI44.2%16.1%†
Code Migration v1.1Vals AI55.1%28.6%†
ProgramBenchVals AI1.5%2.0%
SWE-bench MultimodalVals AI–44.2%†
Vibe Code Bench 1-100Vals AI20.0%18.2%†
Vibe Code Bench v1.1Vals AI78.1%85.0%†
Vibe Code Bench 1-100 v1.1Vals AI29.8%–
SWE Atlas v1.1Scale AI SEAL58.1%†61.0%
SWE Atlas v1.1: Codebase QnAScale AI SEAL70.2%†65.3%
SWE Atlas v1.1: Test WritingScale AI SEAL55.6%†63.3%
SWE Atlas v1.1: RefactoringScale AI SEAL48.6%†54.3%
SWE-Bench Pro V2Scale AI SEAL95.6%97.7%†
SWE-Bench Pro V2 HARDScale AI SEAL84.3%88.2%†
MirrorCodeEpoch AI Benchmarking Hub–54.4%†
Surface Evolver BenchEpoch AI Benchmarking Hub–95.0%
DeepSWE v1.1Epoch AI Benchmarking Hub69.0%68.5%†
CursorBenchCursor42.6%–
ALE-BenchEpoch AI Benchmarking Hub1317.4†1524.5†
FrontierSWEEpoch AI Benchmarking Hub30.2%25.9%†
WeirdML v3WeirdML–7.5%†
Terminal-Bench 4.0 (AA)Artificial Analysis41.9%12.6%†
Terminal-Bench 2.1 (AA)Artificial Analysis83.9%85.0%†
Startrise FrontendStartrise73.2%†79.8%
GameBench 2 OverallDuelLab GameBench 264.6†48.0
GameBench 2 by reasoning settingDuelLab GameBench 272.345.7†
Design Arena WebsiteDesign Arena1279†1336
FrontierCode 1.1 (AA)Artificial Analysis40.1%43.6%†
FrontierCode 2.0 (AA)Artificial Analysis21.6%10.9%†
Agents & tools
Terminal-BenchTerminal-Bench41.2%23.0%†
APEX-AgentsMercor56.6%†47.8%
Arena AgentArena (LMArena)10.6†
LiveBench Long AgenticLiveBench–36.0%†
LiveBench BrowserLiveBench–75.2%†
LiveBench Agentic CodingLiveBench60.9%†62.2%
Terminal-Bench 2.1 (Vals)Vals AI71.5%80.9%
MCP AtlasScale AI SEAL84.2%†82.3%†
HiL-BenchScale AI SEAL58.3%†42.7%
τ³-bench Bankingτ²-bench–37.1%†
GDPval-AA v2.1, normalised scoreArtificial Analysis57.5%–
GDPval-AA v2.1Artificial Analysis16551542†
τ³-Banking (AA)Artificial Analysis50.3%46.0%†
ITBench SRE (AA)Artificial Analysis46.1%47.7%†
Analyst Agent (AA)Artificial Analysis38.8%38.8%†
APEX-Agents (AA)Artificial Analysis–41.3%†
CyberBenchVals AI72.1%75.2%†
Finance Agent v2Vals AI55.8%53.1%†
Finance Agent v2.1Vals AI65.0%60.0%†
Harvey's Legal Agent Benchmark v1.1Vals AI14.6%12.9%†
Harvey's Legal Agent Benchmark v1Vals AI8.3%12.9%†
Legal Research BenchVals AI49.0%46.1%†
SkillsBenchVals AI47.5%–
SREBenchVals AI9.2%–
Tax Agent BenchVals AI43.3%33.4%†
Terminal-Bench 4.0 (Vals)Vals AI38.9%17.2%†
Terminal-Bench ScienceVals AI5.7%1.4%†
Time Horizon Index: KSPVals AI–10.5%†
Arena Agent Task CompletionArena (LMArena)3.13.4†
PostTrainBenchEpoch AI Benchmarking Hub34.5%31.5%
GBAEvalEpoch AI Benchmarking Hub–48.4%
Vending-Bench 2Andon Labs8163.6†5165.0
AutomationBenchArtificial Analysis62.2%58.3%†
GDP.pdf (AA)Artificial Analysis11.2%22.0%†
Harvey LAB-AA v1.1Harvey0.3%5.3%†
EnterpriseOps-GymArtificial Analysis36.4%45.3%†
AA-Briefcase v1.1Artificial Analysis15011497†
GDPxGDPx by SemiAnalysis2.7%3.7%†
Design Arena AgenticDesign Arena1239†1261
Terminal-Bench-Science 0.1 (AA)Artificial Analysis9.5%8.1%†
CWE-Bench (AA)Artificial Analysis55.8%53.3%†
DeepSec-Bench (AA)Artificial Analysis24.3%12.4%†
CyberGym end-to-end (AA)Artificial Analysis29.0%58.0%†
FACTS SearchGoogle87.7%†90.3%
Maths
FrontierMath Tiers 1–3Epoch AI Benchmarking Hub68.8%72.2%†
FrontierMath Tier 4Epoch AI Benchmarking Hub29.3%39.0%†
OTIS Mock AIMEEpoch AI Benchmarking Hub91.1%97.2%†
ProofBenchVals AI49.0%87.0%
LiveBench MathematicsLiveBench87.9%†84.4%
AIME 2026MathArena–96.7%†
HMMT February 2026MathArena–97.0%†
MathArena ApexMathArena–65.6%†
Arena MathsArena (LMArena)15011502†
FrontierMath ErdősEpoch AI Benchmarking Hub–0.0%†
Knowledge
SimpleQA VerifiedEpoch AI Benchmarking Hub41.0%50.6%†
LiveBench 2.0 Data AnalysisLiveBench–78.7%†
LiveBench ResearchLiveBench–79.8%†
LiveBench 1.0 Data AnalysisLiveBench70.2%†78.7%
AA-Omniscience v1.1Artificial Analysis6.0-1.7†
MMLU-ProVals AI86.8%88.0%†
LegalBenchVals AI84.8%86.2%†
CorpFinVals AI–71.6%
TaxEvalVals AI72.4%75.7%
BioMysteryBenchVals AI–72.6%†
Excel Modeling BenchmarkVals AI56.3%66.7%†
Excel Modeling Benchmark v1.1Vals AI69.1%71.3%†
MedCodeVals AI42.9%49.4%†
MedScribeVals AI88.8%88.0%†
Public Benefits Bench v1.1Vals AI68.5%68.2%†
Access to Justice BenchmarkVals AI73.7%68.2%†
Financial Coaching BenchVals AI74.1%74.5%†
Immigration BenchVals AI66.7%68.3%†
GDP.pdfEpoch AI Benchmarking Hub–19.0%†
AA-Omniscience v1.1: accuracyArtificial Analysis32.3%42.4%†
AA-Omniscience v1.1: non-hallucinationArtificial Analysis61.1%23.4%†
HealthBench Professional (AA)OpenAI52.8%53.2%†
FACTS ParametricGoogle40.3%†–
Instruction following
LiveBench LanguageLiveBench79.9%†85.5%
LiveBench Instruction FollowingLiveBench69.3%†71.4%
Arena Instruction FollowingArena (LMArena)14811489†
Human preference
Arena TextArena (LMArena)14821487†
EQ-Bench 4EQ-Bench–1319
EQ-Bench Creative Writing v3EQ-Bench2076†2083
Arena Creative WritingArena (LMArena)14561459†
Arena Multi-turnArena (LMArena)14821502†
Arena Longer QueriesArena (LMArena)14881500†
Multimodal
LiveBench DesignLiveBench–72.5%†
MMMU-ProArtificial Analysis–80.5%†
MMMU-Pro (Vals)Vals AI–88.2%
MortgageTaxVals AI–66.3%
SAGEVals AI–86.8%†
Humanity's Sixth SenseScale AI SEAL–25.5%†
Blueprint-Bench 2Epoch AI Benchmarking Hub–29.5%
Furniture AssemblyEpoch AI Benchmarking Hub–34.2%†
Code2Video-AAArtificial Analysis10341041†
AA-3DArtificial Analysis794984†
Long context
AA-LCR v1.1Artificial Analysis79.7%88.7%†
MLCRArtificial Analysis17.2%21.7%†
FACTS GroundingGoogle55.2%†–
Composite
Epoch Capabilities IndexEpoch AI Benchmarking Hub157.1†158.6
LiveBench 2.0LiveBench–65.0%†
LiveBench 1.0LiveBench76.1%†79.2%
AA Intelligence Index v5.0.1Artificial Analysis31.617.4†
AA Intelligence Index v4.3.2Artificial Analysis44.843.6†
AA Progress Index v1Artificial Analysis322.7319.9†
Vals IndexVals AI59.953.0†
Vals Finance IndexVals AI61.4%58.0%†
Vals RSI IndexVals AI22.9%20.9%†
Vals Multimodal IndexVals AI–73.4%
Vals Router IndexVals AI59.7%†53.0%
Vals Legal IndexVals AI74.6%71.7%†
AA Cyber IndexArtificial Analysis36.441.3†
Safety & honesty
FORTRESSScale AI SEAL28.2%27.3%
DistressBenchScale AI SEAL72.5%†73.5%
Metadata
AA Openness IndexArtificial Analysis33.338.9†

Data as of 2026-11-16. Best configuration of each model; every score links to its source on the model pages.

DuelLab GameBench 2, release gb2-20261111-7b598bb13de4 (https://benchmarks.duellab.org/releases/gb2-20261111-7b598bb13de4/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

What changed

  • Kimi K3: BenchLeader Index 59.1 → 60.7.
  • GLM 5.3 (max): 1 new result (PostTrainBench).
  • Kimi K3: 1 score revision.
  • GLM 5.3 (max): 4 score revisions.
  • Kimi K3: 4 score revisions.
  • Kimi K3: 1 score revision.
  • Kimi K3: 1 new result (FACTS Search).
  • GLM 5.3 (max): 1 score revision; cached input price moved from $0.30/M to $0.26/M (13% down).
  • Kimi K3: 1 score revision.
  • GLM 5.3 (max): cached input price moved from $0.26/M to $0.30/M (15% up).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs

GLM 5.3 vs Kimi K3: questions

Is GLM 5.3 better than Kimi K3?
GLM 5.3 (max) and Kimi K3 are neck and neck on quality (58.8 vs 60.7, within the ±6.7 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 16 Nov 2026, so check the category scores for your use.
GLM 5.3 vs Kimi K3: which is cheaper, and will that change?
GLM 5.3 is cheaper today: $2.15 against $6.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
Is GLM 5.3 or Kimi K3 better for coding?
Kimi K3 scores higher in coding (62 vs 60 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 59.0% for GLM 5.3 against 58.7% for Kimi K3.
Which is faster, GLM 5.3 or Kimi K3?
GLM 5.3 streams faster (88 vs 76 output tokens per second, Artificial Analysis 7-day median).
Is GLM 5.3 or Kimi K3 better for agents and tool use?
GLM 5.3 scores higher in agentic tasks (58 vs 54 on the category index, where 50 is average). On Terminal-Bench 2.1 (Vals), the top shared agents & tools benchmark, it is 71.5% for GLM 5.3 against 80.9% for Kimi K3.
More questions (4)
Which is better for long documents?
Both take about 1M tokens: GLM 5.3 1,000,000, Kimi K3 1,048,576.
Which has lower latency?
Kimi K3 answers first: 0.8 s against 57.6 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.
Do independent benchmarks agree on GLM 5.3 vs Kimi K3?
Across the 10 benchmarks both have, GLM 5.3 leads 3, Kimi K3 leads 7. Independent sources split: GLM 5.3 leads on AA Intelligence Index v5.0.1 and Vals Index; Kimi K3 leads on Arena Text.
When was GLM 5.3 released?
GLM 5.3 was released on 18 Aug 2026.