Z.ai, the developer of the GLM models, sells GLM-5.3 on its pay-as-you-go API and through a separate GLM Coding Plan subscription. This page covers the per-token rates, workload costs, and the cheaper and free GLM models worth knowing about.
GLM-5.3 is $1.40 per 1M input tokens, $0.26 per 1M cached input tokens and $4.40 per 1M output tokens, with a 1M context window. It replaced GLM-5.2 at the same price, and storage for cached input is free for a limited time.
| Model | Tier | Input / 1M | Output / 1M | Cache hit / 1M | Context | Notes |
|---|---|---|---|---|---|---|
| GLM-5.3 | Mid-tier | $1.40 | $4.40 | $0.26 | 1M | GLM flagship; replaced 5.2 at same price; strong coding value |
| Model | Support chatbot 10,000 conversations × (2,000 input + 400 output tokens) | Document Q&A 1,000 questions × (30,000 input + 800 output tokens) | Coding agent session 4M input tokens (90% cache hits) + 150,000 output tokens |
|---|---|---|---|
| GLM-5.3 | $45.60 | $45.52 | $2.16 |
Standard rates per 1M tokens from data on this site. The coding-agent row assumes 90% of input is read from the prompt cache (full input price where a model has no cache rate). Cache-write fees, batch discounts and long-context surcharges listed in the notes are not included. Updated November 2, 2026.
| Model | Provider | Tier | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|---|
| GLM-5.3 | Z.ai | Mid-tier | $1.40 | $4.40 | 1M |
| Mistral Large 3 | Mistral | Mid-tier | $0.50 | $1.50 | 256K |
| DeepSeek V4-Pro | DeepSeek | Mid-tier | $0.66 | $1.98 | 1M |
| Mistral Large 4 | Mistral | Mid-tier | $0.70 | $2.10 | 1M |
| Grok 4.3 | SpaceXAI | Mid-tier | $1.25 | $2.50 | 1M |
Z.ai's cheapest model per tier (bold) against the four cheapest same-tier models from other providers, by output price.
GLM-5.3 no longer lets you switch thinking off; instead you set a reasoning effort of low, high or max, and the default is max. For classification, extraction and short chat replies, test low effort first and compare output token counts against max before going to production — at $4.40 per 1M output tokens, the setting is your main lever on cost. GLM-5.3 accepts text only, with no image input, and can return up to 128K output tokens per request.
Z.ai's price list also includes GLM-5.3-Flash and GLM-5.3-FlashX, lighter siblings at a fraction of GLM-5.3's rates, plus three models that are free to call: GLM-4.7-Flash, GLM-4.5-Flash and the GLM-4.6V-Flash vision model. GLM-5.2 and GLM-5.1 are still listed at the same price as GLM-5.3, so staying on them saves nothing.
If you mostly use GLM inside coding agents, the GLM Coding Plan is a subscription (Individual and Team plans) that includes GLM-5.3 with a points-based quota; calls in off-peak hours, including all day at weekends, use half the usual points. Pay-as-you-go API billing is the better fit for production apps and variable traffic. Z.ai notes that Coding Plan subscribers, including those whose plans have expired, can currently reach GLM-5.3 only through its OpenAI Chat Completions-compatible protocol.
GLM-5.3 weights are published on Hugging Face (zai-org/GLM-5.3, a 753B-parameter model, with a BF16 variant), so you can self-host or buy inference elsewhere — Perplexity's Agent API, for example, lists glm-5.3 at the same $1.40 / $4.40 rates. Among same-tier models, GLM-5.3's output price is close to Meta's Muse Spark 1.3 ($4.25) and well above DeepSeek V4-Pro ($1.98).
GLM-5.3 is $1.40 per 1M input tokens, $0.26 per 1M cached input tokens and $4.40 per 1M output tokens on Z.ai's pay-as-you-go API. Cached-input storage is free for a limited time.
No. Z.ai lists GLM-5.3, GLM-5.2 and GLM-5.1 at identical prices, and GLM-5.3 is built on the same base model as GLM-5.2 with further post-training.
Yes. GLM-4.7-Flash, GLM-4.5-Flash and GLM-4.6V-Flash are listed as free for input, cached input and output.
No. Reasoning is always on; you can only lower the effort to low or high from the default max.
Yes. Z.ai publishes GLM-5.3 on Hugging Face as zai-org/GLM-5.3, alongside a BF16 version and the smaller GLM-5.3-Flash.
How we choose: prices come from each vendor's official pricing page and are re-checked every week. We take no affiliate fees, so rankings aren't for sale.