← Back to Blog
New model

GLM 5.3 Flash API: a 1.3M-token context on FastMetal

FastMetal

Z.ai's GLM 5.3 Flash is available on FastMetal as glm-5.3-flash. It is the lighter version of GLM 5.3, which we already serve, and it takes a longer context than the full model does.

What you get

  • A 1,310,720-token context, the longest of any model we serve. GLM 5.3 itself takes 1,048,576.
  • A low per-token rate, well under GLM 5.3's, so feeding it long documents again and again stays affordable. The model page has the current rates.
  • The same family as GLM 5.3. Prompts written for one carry over to the other with little or no change.
  • Image input, tool calling (tools, including a forced tool_choice) and strict json_schema output. We checked each with a real request.

One limit to plan for: requests to glm-5.3-flash are served by two hosts, and one of them accepts up to 1,000,000 tokens. A prompt between 1,000,000 and 1,310,720 tokens can fail with a 400. Keep prompts under 1,000,000 tokens if you need every request to succeed. File content parts (PDF input) are refused on this model; send the text instead.

Calling it

Set model to glm-5.3-flash. The base URL and API key are the same as for every other model on FastMetal:

curl https://api.fastmetal.ai/v1/chat/completions \
  -H "Authorization: Bearer $FASTMETAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "messages": [{"role": "user", "content": "Summarize these meeting notes in 300 words."}]
  }'

With the OpenAI SDK, point base_url at https://api.fastmetal.ai/v1 and change model. To try it without code, pick it in chat.

GLM 5.3 or GLM 5.3 Flash?

Use GLM 5.3 when answer quality matters most, and GLM 5.3 Flash for volume: long inputs, many calls, or both. GLM 5.3 vs GLM 5.3 Flash shows both models' real answers to the same prompts, and the two model pages show the rates side by side.

For another option in the same weight class, see DeepSeek V4.1 Flash vs GLM 5.3 Flash. If your data has to stay in Japan, there is also a Japan-hosted route for this model.

Pricing

Billed per million input and output tokens, with cached input at its own rate, from a prepaid balance with no monthly fee. Current rates are on pricing and the model page.

Try GLM 5.3 Flash right now

GLM 5.3 Flash is available on FastMetal through one API key. Start in the browser, or call it from the OpenAI SDK.