Zhipu AI model product
GLM 5.3
AI's flagship post-trained model for complex coding, long-horizon agentic work, and cybersecurity tasks.
Updated Aug 17, 2026. Default version: GLM-5.3
LLMBoard Score90.9GLM-5.3
Coverage40%16 benchmark families
Context window1MTokens
Official input priceN/AOfficial price unavailable
GLM 5.3 Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
GLM 5.3 Benchmark Results
Benchmark scores for GLM-5.3.
Sort by
Benchmark | Score | Rank | Participants | Percentile | Evidence | Evaluated |
|---|
| BenchmarkAutomationBench | Score48.2% | Rank01 | Participants12 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkCyberGym | Score84.5% | Rank01 | Participants13 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkGDPval-AA | Score1,769 points | Rank01 | Participants4 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkPostTrainBench | Score39.8% | Rank01 | Participants6 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkSWE-Marathon | Score42.5% | Rank01 | Participants4 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench 3.0 | Score28.3% | Rank01 | Participants3 | Percentile100.0% | EvidenceC | Evaluated |
| BenchmarkNL2Repo | Score58.0% | Rank02 | Participants17 | Percentile93.8% | EvidenceC | Evaluated |
| BenchmarkExploitBench | Score54.4% | Rank03 | Participants5 | Percentile50.0% | EvidenceC | Evaluated |
| BenchmarkExploitGym | Score15.0% | Rank03 | Participants4 | Percentile33.3% | EvidenceC | Evaluated |
| BenchmarkFrontierSWE | Score78.1% | Rank03 | Participants16 | Percentile86.7% | EvidenceC | Evaluated |
| BenchmarkTerminal-Bench 2.1 | Score88.2% | Rank03 | Participants28 | Percentile92.6% | EvidenceC | Evaluated |
| BenchmarkHumanity's Last Exam | Score62.5% | Rank04 | Participants99 | Percentile96.9% | EvidenceC | Evaluated |
| BenchmarkToolathlon | Score73.0% | Rank04 | Participants37 | Percentile91.7% | EvidenceC | Evaluated |
| BenchmarkProgram Bench | Score19.0% | Rank06 | Participants6 | Percentile0.0% | EvidenceC | Evaluated |
| BenchmarkAgents' Last Exam | Score28.5% | Rank07 | Participants10 | Percentile33.3% | EvidenceC | Evaluated |
| BenchmarkDeepSWE 1.1 | Score66.9% | Rank08 | Participants25 | Percentile70.8% | EvidenceC | Evaluated |
GLM 5.3 Arena Results
Preference and agent-evaluation results for the default version.
No Arena results
The default version does not have a matching Arena result yet.
GLM 5.3 Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $1.4 input, $4.4 output per 1M via OpenCode Go
- Tracked offerings
- 3
Sort by
Provider | Provider model ID | Region | Input / 1M | Output / 1M | Context | Updated |
|---|
| ProviderZhipu AI Coding Plan | Provider model IDglm-5.3 | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context1M | Updated |
| ProviderZ.AI Coding Plan | Provider model IDglm-5.3 | Regionglobal | Input / 1MN/A | Output / 1MN/A | Context1M | Updated |
| ProviderOpenCode Go | Provider model IDglm-5.3 | Regionglobal | Input / 1M$1.4 | Output / 1M$4.4 | Context1M | Updated |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
GLM 5.3 Runtime Performance
Provider-specific output speed and catalog latency for GLM-5.3. Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
GLM 5.3 Specifications
Technical details for the model's default version.
- Version
- GLM-5.3
- Released
- Aug 14, 2026
- Knowledge cutoff
- Unknown
- Parameters
- 753B
- Context window
- 1M
- Max output
- 131.1K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Unknown
GLM 5.3 Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
Sort by
Version | Released | LLMBoard | Parameters | Context | Max output | Open weights | License |
|---|
| VersionGLM-5.3 | Released | LLMBoard90.9 | Parameters753B | Context1M | Max output131.1K | Open weightsNo | LicenseUnknown |
GLM 5.3 vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to GLM 5.3
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
ZA
GLM 5.2
Zhipu AI
ZA
GLM 5.1
Zhipu AI
ZA
GLM 5
Zhipu AI
ZA
GLM 4.7
Zhipu AI
ZA
GLM 5V Turbo
Zhipu AI
ZA
GLM 4.6
Zhipu AI
What is GLM 5.3?
Key information about GLM 5.3 and its available data.
AI's flagship post-trained model for complex coding, long-horizon agentic work, and cybersecurity tasks. 2, with its capability gains coming from post-training. The model supports a 1M-token context window and always-on thinking with low, high, and max reasoning-effort levels.
3 through ZCode and the GLM Coding Plan on August 14, 2026, and announced that open weights would follow after a two-week safety evaluation and hardening period.
Data as of 2026-08-17.
FAQ
Common questions about GLM 5.3.
When was GLM 5.3 released?
GLM 5.3's default version was released on Aug 14, 2026.
How much does GLM 5.3 cost?
No official standard PAYG price is currently available for GLM 5.3. The lowest tracked third-party offer starts at $1.4 input and $4.4 output via OpenCode Go.
Who created GLM 5.3?
GLM 5.3 was created by Zhipu AI.
What is the context window for GLM 5.3?
The default version has a 1M token context window.
Is GLM 5.3 open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer GLM 5.3?
3 provider offerings are linked to the default version.
What models should I compare GLM 5.3 with?
Nearby ranked alternatives include Claude Fable 5, Claude Mythos, Kimi K3.