GLM 5.3 Flash
$0.041 per 1M input tokens and $0.12 per 1M output tokens · 1.05M context · vision, tools, reasoning, JSON, open weights.
Prices checked Oct 21, 2026 · tracked by LLM Cost Index since Oct 21, 2026
What GLM 5.3 Flash costs per month
Estimated monthly bill for typical workloads, using the cached-input price for repeated context. Try your own numbers in the cost calculator.
| Workload | Per month |
|---|---|
| Customer chatbot 1,000 conversations a day, 8 turns each | $16.82 |
| Document Q&A (RAG) 500 questions a day over retrieved documents | $4.27 |
| Coding / tool-using agent 200 tasks a day, long tool-heavy contexts | $7.60 |
| Content generation 30 long articles a day | $0.381 |
| Classification / extraction 100,000 short items a day | $42.47 |
Price history
We started tracking GLM 5.3 Flash on Oct 21, 2026; the chart appears as soon as there is a second data point. See all recent price changes.
Where to run GLM 5.3 Flash: 35 providers
OpenInference is the cheapest at $0.041 input / $0.12 output per 1M tokens. Uptime is the average over the last 30 days as reported by OpenRouter; see provider reliability.
| Provider | Input / 1M | Output / 1M | Context | Uptime | Notes |
|---|---|---|---|---|---|
| OpenInference | $0.041 | $0.12 | 1.05M | 97.8% | fp4 |
| StreamLake | $0.059 | $0.196 | 1.02M | 92.3% | fp8 |
| Wafer | $0.074 | $0.20 | 1.05M | 97.4% | nvfp4 |
| GMICloud | $0.075 | $0.25 | 1.05M | 93.5% | fp8 |
| Relace | $0.075 | $0.16 | 1.05M | 99% | — |
| Near AI | $0.08 | $0.50 | 1.05M | 99.2% | fp8 |
| Novita | $0.084 | $0.28 | 1.05M | 94.8% | fp8 |
| Sail Research | $0.09 | $0.15 | 1.05M | 100% | fp4 |
| Sail Research | $0.09 | $0.15 | 1.05M | 100% | fp4 |
| Reka | $0.10 | $2.40 | 1.05M | 92.5% | — |
| InferenceNet | $0.12 | $0.388 | 1.05M | 94.4% | fp4 |
| InferenceNet | $0.12 | $0.388 | 1.05M | 94% | fp4 |
| NextBit | $0.12 | $0.60 | 1.05M | 76.8% | fp4 |
| Phala | $0.128 | $0.425 | 1.05M | 94.2% | nvfp4 |
| DigitalOcean | $0.131 | $0.438 | 1.05M | 94.7% | — |
| Makora | $0.14 | $0.47 | 1.05M | 65.9% | — |
| AtlasCloud | $0.15 | $0.50 | 1.05M | 94.6% | fp8 |
| BaseTen | $0.15 | $0.50 | 1.05M | 94.7% | fp8 |
| BaseTen | $0.15 | $0.50 | 1.05M | 96.1% | fp8 |
| CoreWeave | $0.15 | $0.50 | 1.05M | 96% | nvfp4 |
| Crusoe | $0.15 | $0.50 | 1.05M | 84% | fp4 |
| DeepInfra | $0.15 | $0.50 | 1.05M | 96.8% | fp4 |
| Fireworks | $0.15 | $0.50 | 1.05M | 92.5% | — |
| Friendli | $0.15 | $0.50 | 1.05M | 98.2% | — |
| Modal | $0.15 | $0.50 | 1.05M | 100% | nvfp4 |
| Parasail | $0.15 | $0.50 | 1.05M | 92.9% | fp4 |
| PrimeIntellect | $0.15 | $0.50 | 1.05M | 99.9% | — |
| SiliconFlow | $0.15 | $0.50 | 1.05M | 93.6% | fp8 |
| Together | $0.15 | $0.50 | 1.05M | 87.2% | — |
| Venice | $0.15 | $0.50 | 1.05M | 96.7% | — |
| Z.AI | $0.15 | $0.50 | 1.05M | 99.7% | fp8 |
| Inceptron | $0.175 | $0.69 | 1.05M | 96.3% | fp8 |
| Parasail | $0.188 | $0.625 | 1.05M | 96.9% | priority tier, fp4 |
| Fireworks | $0.225 | $0.75 | 1.05M | 100% | — |
| Morph | $0.54 | $0.65 | 1.05M | 95.6% | fp8 |
How to pay less for GLM 5.3 Flash
For a customer chatbot (1,000 conversations a day, 8 turns each) that costs $24.91 a month at list price without caching:
| Option | Per month | You save |
|---|---|---|
| Use prompt caching (60% of input cached) | $16.82 | $8.09 (32%) |
Other workloads: cut-your-bill tool for GLM 5.3 Flash.
Compare GLM 5.3 Flash
Specifications
| Model ID | z-ai/glm-5.3-flash |
| Input | text, image, video |
| Output | text |
| Tool calling | Yes |
| Structured output | Yes |
| Reasoning mode | Yes |
| Open weights | Yes — zai-org/GLM-5.3-Flash |
| Released | Aug 26, 2026 |
FAQ
- How much does GLM 5.3 Flash cost?
- $0.041 per million input tokens and $0.12 per million output tokens; cached input costs $0.0037 per million. Prices as of Oct 21, 2026.
- What is the context window of GLM 5.3 Flash?
- 1.05M tokens, with up to 944K output tokens per response.
- Does GLM 5.3 Flash support images, tools and structured output?
- Images as input: yes. Tool calling: yes. Structured (JSON) output: yes. Reasoning mode: yes.
More from Z.ai: all Z.ai models · how we collect prices