GLM-5.3-Flash
Zhipu AI
Key facts · verified Oct 30, 2026
GLM-5.3-Flash is a Zhipu AI model released on 26 Aug 2026. Its API price is $0.15 per million input tokens and $0.50 per million output tokens ($0.15 input / $0.50 output per 1M tokens ($0.03 cached input)). Its context window is 1M tokens. It ranks #33 on LMArena's text leaderboard (Elo 1477, read 2 Nov 2026). Source: docs.z.ai.
Z.ai's GLM-5.3-Flash, released 26 August 2026 and the first natively multimodal model in the GLM-5 series. A 320B mixture-of-experts with 18B active parameters, a 1M-token context window and 128k maximum output, published on Hugging Face under the MIT licence. It ran anonymously as “Ox Alpha” on OpenCode and OpenRouter for its first week. List pricing is roughly a tenth of GLM-5.3.
Pricing & specs checkedSource: vendor documentation · checked Oct 30, 2026LMArena: 1477 Elo (rank 33) · source · retrieved Nov 2, 2026
Key Features
Native Multimodal Agentic Coding Long Context Thinking Mode Context Caching Function Calling Structured Output
Full Specifications
$0.15 / $0.50 per million tokens | |
|---|---|
| Identity | |
| Developer | Zhipu AI |
| Released | Aug 2026 |
| Status | GA |
| Licence | MIT (weights on Hugging Face, zai-org/GLM-5.3-Flash) |
| Self-hostable | Yes - MIT weights published at launch |
| Cost | |
| Blended $/1M tokensinput × 0.75 + output × 0.25 | $0.237 / 1M tokens |
| Input price | $0.150 / 1M tokens |
| Output price | $0.500 / 1M tokens |
| Cached input | 0.03 |
| Batch discount | Not offered; GLM Coding Plan applies a 50% off-peak points reduction |
| Free tier | Not offered; included in the GLM Coding Plan at 3x the GLM-5.3 quota |
| Capacity | |
| Context window | 1M tokens |
| Max output | 128k tokens |
| Long-context surcharge | Not published |
| Capability | |
| Vision in | Yes - text, image, video and file input |
| Audio in | Not documented |
| Function calling | Yes |
| Structured output | Yes |
| Extended reasoning | Yes - thinking mode |
| Web search | Not documented |
| Code execution | Not documented |
| Access | |
| API | Yes |
| Chat app | Yes |
| Cloud marketplaces | Not documented |
| Fine-tuning | Not offered |
| Measured quality | |
| LMArena Elo | 1477 (checked Nov 2026) |
| LMArena rank | Rank 33 |
| GPQA Diamond | 90 · max effort |
| Elo per dollarLMArena Elo ÷ blended $/1M tokens | 6219 |