AIEvals

Models

GLM 5.3 Flash

Specifications and API prices for GLM 5.3 Flash, recorded by Artificial Analysis.

Z.aiReleased 26 August 2026Context 1M tokensOutput Not published in the checked sourceOpen weights$0.15 in / $0.50 out per Mtok

Family GLM38 benchmarks with a published resultNo first place on the independent columns among the 46 models this site listsRead 17 October 2026

Where it stands

Rank compares tracked models with a score. Published shows the source’s rank and field size, when available.

BenchmarkScoreRankPublishedLeaderBehind
Artificial Analysis Intelligence IndexArtificial Analysis4220of 42no list readClaude Opus 5.516 pts
Terminal-Bench 4.0Artificial Analysis32.83%17of 40no list readClaude Sonnet 5.530.81 pts
SWE-bench (Verified)Vals AI92.00%10of 2310of 88Claude Opus 55.00 pts
LiveCodeBenchVals AI80.51%21of 2169of 143Claude Fable 5.110.01 pts
GPQA DiamondVals AI86.36%22of 2343of 138Gemini 3.1 Pro9.09 pts
MMLU ProVals AI86.06%21of 2449of 138Claude Fable 5.16.32 pts
MMMU ProVals AI86.01%14of 2023of 93Claude Fable 5.14.63 pts
Vibe Code Bench v1.1Vals AI30.76%36of 3865of 106Claude Sonnet 5.561.63 pts
Terminal-Bench Science 0.1Artificial Analysis4.76%20of 24no list readGPT-6 Astra58.57 pts
ProgramBenchVals AI0.00%19of 2923of 54Claude Opus 5.518.50 pts
τ³-BankingArtificial Analysis47.22%6of 30no list readQwen3.8 Max4.12 pts
ITBench-AAArtificial Analysis51.24%3of 18no list readGPT-5.6 Sol4.97 pts
MLCR-AAArtificial Analysis51.11%7of 37no list readClaude Sonnet 5.523.89 pts
Vibe Code Bench 1–100Vals AI16.03%13of 2013of 20Claude Opus 5.514.33 pts
EnterpriseOps-Gym-AAArtificial Analysis33.21%21of 22no list readClaude Fable 517.91 pts
AutomationBench-AAArtificial Analysis60.37%13of 40no list readGemini 4 Argon17.14 pts
SciCodeArtificial Analysis51.62%34of 40no list readClaude Opus 5.515.28 pts
GDP.pdfArtificial Analysis15.40%30of 40no list readGPT-6.1 Sol15.60 pts
CritPtArtificial Analysis15.43%33of 42no list readGPT-5.6 Sol16.86 pts
AA-OmniscienceArtificial Analysis7.4729of 42no list readClaude Opus 5.538.95
AA-LCR v1.1Artificial Analysis80.00%29of 42no list readKimi K38.67 pts
Terminal-Bench 2.1Vals AI62.92%30of 3538of 75Claude Opus 5.524.72 pts
ProofBench v1.1Vals AI21.00%31of 3437of 46Claude Sonnet 5.579.00 pts
LegalBenchVals AI83.93%20of 2743of 149Claude Fable 54.63 pts
Harvey's Legal Agent BenchmarkVals AI6.67%18of 3724of 73Muse Spark 1.218.75 pts
SkillsBenchVals AI40.16%19of 2030of 35DeepSeek V4.1 Flash29.64 pts
Code MigrationVals AI20.52%32of 3745of 71Claude Sonnet 5.549.31 pts
CyberBench v1.1Vals AI67.86%18of 3123of 43GPT-6 Sol10.12 pts
EMBVals AI55.93%31of 3742of 69Claude Fable 5.120.74 pts
Finance Agent (v2)Vals AI57.85%10of 3712of 73Gemini 4 Argon7.55 pts
IOIVals AI52.50%20of 3121of 38Gemini 4 Argon47.50 pts
Legal Research BenchVals AI45.19%14of 3714of 72Muse Spark 1.3 Max10.10 pts
MedScribeVals AI88.94%7of 367of 106Claude Opus 5.52.49 pts
Tax Agent BenchVals AI61.96%25of 3729of 64Claude Fable 5.115.68 pts
DeepSWE v1.1Datacurve63.39%13of 2013of 28GPT-6 Astra10.73 pts
GDPval-AA v2.1Artificial Analysis1,640.7810of 42no list readClaude Opus 5.5205.39
AA-Briefcase v1.1Artificial Analysis1,452.4621of 41no list readClaude Opus 5.5369.39
Humanity's Last ExamArtificial Analysis39.85%34of 42no list readClaude Opus 5.521.50 pts

Nearby on Artificial Analysis Intelligence Index

The two models above and below on the Artificial Analysis Intelligence Index (Artificial Analysis), on every column where this model has a published number.

ModelArtificial Analysis Intelligence IndexindexTerminal-Bench 4.0%SWE-bench%LiveCodeBench%GPQA Diamond%MMLU Pro%MMMU Pro%Vibe Code Bench v1.1%Terminal-Bench Science 0.1%ProgramBench%
Claude Haiku 5.5Anthropic4332.83%––––––20.00%–
Claude Opus 4.8Anthropic4221.72%88.60%87.82%92.42%89.59%86.59%82.72%–1.00%
GLM 5.3 FlashZ.ai4232.83%92.00%80.51%86.36%86.06%86.01%30.76%4.76%0.00%
GPT-5.6 TerraOpenAI4235.35%95.40%85.93%90.91%86.66%86.47%74.59%13.81%0.50%
Claude Opus 4.7Anthropic41–82.00%85.07%90.15%89.87%85.55%71.00%–0.00%
Bold marks a figure better than this model’s on the same column, whichever model it belongs to. Ranks cover the models shown from the stated source.

Cost and speed, as published

Additional measurements from Artificial Analysis’ own row. Read 30 September 2026.

As publishedCost per taskOutput speed
GLM 5.3 Flash$0.2545 tok/s

Results by benchmark

Independent and vendor results retain their own harnesses and effort settings.

BenchmarkIndependent runAs reported by a publisher
τ³-BankingTool use47.22%Artificial Analysisnot published
ITBench-AAIT operations51.24%Artificial Analysisnot published
MLCR-AAMedical reasoning51.11%Artificial Analysisnot published
Vibe Code Bench 1–100Application development16.03%Vals AInot published
EnterpriseOps-Gym-AAEnterprise workflows33.21%Artificial Analysisnot published
AutomationBench-AAWorkflow automation60.37%Artificial Analysisnot published
SciCodeScientific coding51.62%Artificial Analysisnot published
GDP.pdfDocument reasoning15.40%Artificial Analysisnot published
CritPtScientific reasoning15.43%Artificial Analysisnot published
AA-OmniscienceFactual knowledge7.47Artificial Analysisnot published
AA-LCR v1.1Long-context reasoning80.00%Artificial Analysisnot published
Terminal-Bench 4.0Agentic coding32.83%Artificial Analysisnot published
Terminal-Bench 2.1Agentic coding62.92%Vals AI
84.27%Artificial Analysis
not published
SWE-bench (Verified)Software engineering92.00%Vals AInot published
LiveCodeBenchCode generation80.51%Vals AInot published
GPQA DiamondScience reasoning86.36%Vals AI
91.21%Artificial Analysis
not published
MMLU ProKnowledge86.06%Vals AInot published
MMMU ProMultimodal reasoning86.01%Vals AInot published
Vibe Code Bench v1.1Product coding30.76%Vals AInot published
Terminal-Bench Science 0.1Agentic science4.76%Artificial Analysis
2.86%Vals AI
not published
ProgramBenchAgentic coding0.00%Vals AInot published
ProofBench v1.1Mathematics21.00%Vals AInot published
LegalBenchProfessional knowledge83.93%Vals AInot published
Harvey's Legal Agent BenchmarkProfessional knowledge6.67%Vals AInot published
SkillsBenchAgentic tooling40.16%Vals AInot published
Code MigrationAgentic coding20.52%Vals AInot published
CyberBench v1.1Security67.86%Vals AInot published
EMBFinancial modelling55.93%Vals AInot published
Finance Agent (v2)Financial analysis57.85%Vals AInot published
IOICompetitive programming52.50%Vals AInot published
Legal Research BenchProfessional knowledge45.19%Vals AInot published
MedScribeHealthcare administration88.94%Vals AInot published
Tax Agent BenchProfessional knowledge61.96%Vals AInot published
DeepSWE v1.1Software engineering63.39%Datacurvenot published
GDPval-AA v2.1Knowledge work1,640.78Artificial Analysisnot published
AA-Briefcase v1.1Knowledge work1,452.46Artificial Analysisnot published
Humanity's Last ExamMultidisciplinary reasoning39.85%Artificial Analysisnot published
Artificial Analysis Intelligence IndexComposite42Artificial Analysisnot published
Independent column: the suite named under each figure, one harness per suite. Publisher column: each lab’s own table, reproduced on the benchmark’s page. Publisher ranks compare reported scores in the same table row only; ties share a rank and unreported cells are excluded.

Sources

Read on 17 October 2026.