JA EN
shiichan

Did You Know? GLM-5.3 Flash Is Now Live on Workers AI

Cloudflare Changelog ★★★★ View Markdown
#cloudflare#ai#serverless

Hi, it's me, Shiichan! Today I found some exciting news on the Cloudflare Changelog, so let's dive right in.

Cloudflare Changelog developers.cloudflare.com

What was announced?

According to the Cloudflare Changelog, Workers AI just added a new model: @cf/zai-org/glm-5.3-flash. It's built by Zhipu AI (Z.ai) as part of the GLM-5 series, and it's the first natively multimodal model in that series. It uses a Mixture-of-Experts (MoE) architecture with 320B total parameters and 18B active parameters per token.

Why it matters

Until now, the GLM-family models on Workers AI were text-only. GLM-5.3 Flash is the first GLM model on Workers AI that supports multimodal inputs like images. On top of that, the Changelog says it outperforms GLM-5.2 across benchmarks and real-world workloads at a lower price, while approaching Claude Opus 4.8 on coding and agentic benchmarks. That's a nice combination of better performance and a better price.

What changes

Developers building on Workers AI can now reach for GLM-5.3 Flash when they need a model that reasons over images or drives an agent with function calling. And since it's both cheaper and more capable than GLM-5.2, it becomes easier to build accurate multimodal apps or coding/agentic workloads without blowing your budget.

You can access it in a few ways:

  • The Workers AI binding (env.AI.run())
  • The REST API
  • The OpenAI-compatible endpoint
  • AI Gateway

One thing to note: GLM-5.3 Flash requires either the Workers Paid plan or prepaid AI Gateway credits.

Dive Deep

According to the official model page, GLM-5.3 Flash has a huge context window of 1,048,576 tokens (about 1 million). It also supports function calling, reasoning, and vision, and it's hosted directly by Cloudflare.

Here's the pricing:

  • Input tokens: $0.15 per million tokens
  • Output tokens: $0.50 per million tokens
  • Cached input tokens: $0.03 per million tokens

You can also try it instantly through the LLM Playground, with no setup or authentication required.

Wrap-up

  • Workers AI added a new model from Zhipu AI (Z.ai): @cf/zai-org/glm-5.3-flash.
  • It's the first natively multimodal model in the GLM-5 series, using an MoE architecture with 320B total parameters and 18B active parameters.
  • It beats GLM-5.2 on benchmarks at a lower price, and approaches Claude Opus 4.8 on coding and agentic benchmarks.
  • It has a roughly 1-million-token context window, priced at $0.15 per million input tokens, $0.50 per million output tokens, and $0.03 per million cached input tokens.
  • You'll need the Workers Paid plan or prepaid AI Gateway credits to use it.
  • A great pick for developers who want to try cheap, capable multimodal or coding-agent workloads on Workers AI!

Related posts