Head-to-head comparison
Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.
Specifications
Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.
| Attribute | GLM-5.3 | GLM-5.2 |
|---|---|---|
| Provider | Z.ai | Z.ai |
| Released | Aug 14, 2026 | Jun 16, 2026 |
| Availability | Active | Open Source |
| Model type | Open Weights | Open Weights |
| Parameters | 744B, 40B active | 744B, 40B active |
| Architecture | Mixture of Experts | Mixture of Experts |
| Context window | 1,000,000 tokens | 1,000,000 tokens |
| Knowledge cutoff | Not published | Not published |
| Technical sources | Model, Context | Model, Context |
Pricing
The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.
| Price type | GLM-5.3 | GLM-5.2 |
|---|---|---|
| API input | $1.4 | $1.4 |
| API output | $4.4 | $4.4 |
| Cache read | $0.26 | $0.26 |
| Cache write (5 min.) | Not listed | Not listed |
| Cache write (1 hr.) | Not listed | Not listed |
| Price verified | 08/30/2026 Source | 07/30/2026 Source |
Key to the marks:better valuebetter, but not conclusivetie
Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.
| Model | Provider | Purchase route and tariff | Input per 1M tokens | Output per 1M tokens | Source |
|---|---|---|---|---|---|
| GLM-5.3 | Z.ai | Direct from developerStandard API | $1.4 | $4.4 | Source08/30/2026 |
| AkashML | Via OpenRouterakashml/fp8 | $1.17 | $3.96 | Source09/08/2026 | |
| AtlasCloud | Via OpenRouteratlas-cloud/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fp4 | $1.4 | $4.4 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fp8 | $2.1 | $6.6 | Source09/08/2026 | |
| Cloudflare | Via OpenRoutercloudflare | $1.4 | $4.4 | Source09/08/2026 | |
| Crusoe | Via OpenRoutercrusoe/fp4 | $1.4 | $4.4 | Source09/08/2026 | |
| Decart | Via OpenRouterdecart/fp4 | $1.19 | $3.74 | Source09/08/2026 | |
| DeepInfra | Via OpenRouterdeepinfra/fp4 | $1.2 | $4 | Source09/08/2026 | |
| DigitalOcean | Via OpenRouterdigitalocean | $1.4 | $4.4 | Source09/08/2026 | |
| Fireworks | Via OpenRouterfireworks | $1.4 | $4.4 | Source09/08/2026 | |
| Friendli | Via OpenRouterfriendli | $1.26 | $3.96 | Source09/08/2026 | |
| GMICloud | Via OpenRoutergmicloud/fp8 | Lowest listed price: $1.12 | Lowest listed price: $3.52 | Source09/08/2026 | |
| Inceptron | Via OpenRouterinceptron/fp4 | $1.25 | $4.4 | Source09/08/2026 | |
| Io Net | Via OpenRouterio-net/fp8 | $1.1875 | $4.18 | Source09/08/2026 | |
| Makora | Via OpenRoutermakora/fp4 | $1.35 | $4.4 | Source09/08/2026 | |
| Modal | Via OpenRoutermodal | $1.4 | $4.4 | Source09/08/2026 | |
| Morph | Via OpenRoutermorph/fp4 | $1.25 | $4.4 | Source09/08/2026 | |
| Novita | Via OpenRouternovita/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| Parasail | Via OpenRouterparasail/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| Phala | Via OpenRouterphala | $1.4 | $4.4 | Source09/08/2026 | |
| Reka | Via OpenRouterreka/fp8 | $1.17 | $3.96 | Source09/08/2026 | |
| Sail Research | Via OpenRoutersail-research/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| SiliconFlow | Via OpenRoutersiliconflow/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| StreamLake | Via OpenRouterstreamlake/fp8 | $1.26 | $3.96 | Source09/08/2026 | |
| Together | Via OpenRoutertogether | $1.4 | $4.4 | Source09/08/2026 | |
| Venice | Via OpenRoutervenice | $1.75 | $5.5 | Source09/08/2026 | |
| Wafer | Via OpenRouterwafer | $1.19 | $4.4 | Source09/08/2026 | |
| Z.AI | Via OpenRouterz-ai/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| GLM-5.2 | Z.ai | Direct from developerStandard API | $1.4 | $4.4 | Source07/30/2026 |
| Alibaba | Via OpenRouteralibaba/fast | $2.31 | $7.26 | Source09/08/2026 | |
| Alibaba | Via OpenRouteralibaba/fp8 | $0.966 | $3.036 | Source09/08/2026 | |
| Ambient | Via OpenRouterambient/fp8 | $0.6 | $2 | Source09/08/2026 | |
| AtlasCloud | Via OpenRouteratlas-cloud/fp8 | $0.938 | $2.948 | Source09/08/2026 | |
| Baidu | Via OpenRouterbaidu/fp4 | $2.25 | $7.88 | Source09/08/2026 | |
| Baidu | Via OpenRouterbaidu/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fast#1 | $2.1 | $6.6 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fast#2 | $2.1 | $6.6 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fp8#1 | $1.4 | $4.4 | Source09/08/2026 | |
| BaseTen | Via OpenRouterbaseten/fp8#2 | $1.4 | $4.4 | Source09/08/2026 | |
| Cloudflare | Via OpenRoutercloudflare | $1.4 | $4.4 | Source09/08/2026 | |
| CoreWeave | Via OpenRoutercoreweave/fp4 | $0.76 | $2.42 | Source09/08/2026 | |
| Crusoe | Via OpenRoutercrusoe/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| Decart | Via OpenRouterdecart/fast | $1.89 | $5.94 | Source09/08/2026 | |
| DeepInfra | Via OpenRouterdeepinfra/fp4 | Lowest listed price: $0.4875 | Lowest listed price: $1.56 | Source09/08/2026 | |
| DigitalOcean | Via OpenRouterdigitalocean | $0.7 | $2.2 | Source09/08/2026 | |
| Fireworks | Via OpenRouterfireworks | $1.4 | $4.4 | Source09/08/2026 | |
| Fireworks | Via OpenRouterfireworks/fast | $2.1 | $6.6 | Source09/08/2026 | |
| Fireworks | Via OpenRouterfireworks/fast-us | $2.1 | $6.6 | Source09/08/2026 | |
| Friendli | Via OpenRouterfriendli | $1.4 | $4.4 | Source09/08/2026 | |
| GMICloud | Via OpenRoutergmicloud/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| Inceptron | Via OpenRouterinceptron/fp4 | $1.25 | $2.99 | Source09/08/2026 | |
| Mistral | Via OpenRoutermistral | $1.4 | $4.4 | Source09/08/2026 | |
| Mistral | Via OpenRoutermistral/eu | $1.54 | $4.84 | Source09/08/2026 | |
| Mistral | Via OpenRoutermistral/zdr | $1.4 | $4.4 | Source09/08/2026 | |
| Novita | Via OpenRouternovita/fp8 | $0.6832 | $2.1472 | Source09/08/2026 | |
| Parasail | Via OpenRouterparasail/fp4 | $1.4 | $4.4 | Source09/08/2026 | |
| Phala | Via OpenRouterphala/fp8 | $1.26 | $3 | Source09/08/2026 | |
| SiliconFlow | Via OpenRoutersiliconflow/fp8 | $1.19 | $3.74 | Source09/08/2026 | |
| StreamLake | Via OpenRouterstreamlake/fp8 | $0.6762 | $2.1252 | Source09/08/2026 | |
| Together | Via OpenRoutertogether | $1.4 | $4.4 | Source09/08/2026 | |
| Venice | Via OpenRoutervenice/fp8 | $1.4 | $4.4 | Source09/08/2026 | |
| Z.AI | Via OpenRouterz-ai/fp8 | $1.4 | $4.4 | Source09/08/2026 |
Performance
Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.
Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.
18 matched rows from 18 benchmark families were considered. The radar shows 8 of 8 comparable categories.
Key to the marks:better valuebetter, but not conclusivetie
| Benchmark and task | GLM-5.3 | GLM-5.2 | Source |
|---|---|---|---|
Code MigrationOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner44.222 % | 37.87 % | Vals AISource detailsData as of: 09/03/2026 |
EMBOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | 56.344 % | Winner61.532 % | Vals AISource detailsData as of: 09/03/2026 |
Finance Agent (v2)OverallTest detailsNo documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better | Winner55.84 % | 49.699 % | Vals AISource detailsData as of: 09/03/2026 |
GPQA DiamondOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner88.132 % | 85.606 % | Vals AISource detailsData as of: 09/01/2026 |
Harvey's Legal Agent BenchmarkOverall · Task fully resolvedTest detailsNo documented difference in the test setupVersion: 1Metric: task resolution rateScoring: Higher is better | Winner8.333 % | 7.083 % | Vals AISource detailsData as of: 09/03/2026 |
Legal Research BenchOverall · All-passTest detailsNo documented difference in the test setupVersion: 1Metric: all-pass rateScoring: Higher is better | Winner49.038 % | 31.25 % | Vals AISource detailsData as of: 09/03/2026 |
LegalBenchOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner84.837 % | 84.073 % | Vals AISource detailsData as of: 09/01/2026 |
LiveCodeBench v6OverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner80.535 % | 69.5 % | Vals AISource detailsData as of: 09/01/2026 |
MedCodeOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner42.864 % | 40.771 % | Vals AISource detailsData as of: 09/03/2026 |
MedScribeOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner88.81 % | 83.534 % | Vals AISource detailsData as of: 09/03/2026 |
MMLU-ProOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner86.77 % | 86.714 % | Vals AISource detailsData as of: 09/01/2026 |
ProgramBenchOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner1.5 % | 0.5 % | Vals AISource detailsData as of: 09/01/2026 |
SkillsBenchOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands | Winner47.509 % | 45.077 % | Vals AISource detailsData as of: 09/01/2026 |
SWE-bench VerifiedOverallTest detailsNo documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better | Winner95.4 % | 82.8 % | Vals AISource detailsData as of: 09/01/2026 |
TaxEval v2OverallTest detailsNo documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better | 72.363 % | Winner73.344 % | Vals AISource detailsData as of: 09/01/2026 |
Terminal-Bench 2.1OverallTest detailsNo documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is better | Winner71.536 % | 67.79 % | Vals AISource detailsData as of: 09/03/2026 |
Vals IndexOverallTest detailsNo documented difference in the test setupVersion: 2Metric: weighted index scoreScoring: Higher is better | Winner56.97 % | 53.122 % | Vals AISource detailsData as of: 09/03/2026 |
Vibe Code Bench v1.1OverallTest detailsNo documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands | Winner78.125 % | 63.961 % | Vals AISource detailsData as of: 09/03/2026 |
Key to the marks:better valuebetter, but not conclusivetie | |||
Our model tests
Compare how the models respond to the same prompt. Each test shows the first attempt, with no subsequent fixes to the generated code. These results do not contribute to an overall score.
A community classic for free-form SVG drawing.
GLM-5.3
API access was unavailable for this run. The model was not evaluated.
Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.
GLM-5.2
API access was unavailable for this run. The model was not evaluated.
Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.
Generate an SVG of a pelican riding a bicycle
Token limit including reasoning: 8,192 tokens
Task origin (Simon Willison)How to read this comparison
Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.
The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.
Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.
More matchups
Compare GLM-5.3 and GLM-5.2 with other leading models using the same data and benchmark logic.