cloudflare/ai workers-ai-provider@3.1.12

on GitHub

Patch Changes

• #504 e9b2a9a Thanks @threepointone! - Forward reasoning_effort and chat_template_kwargs onto binding.run(model, inputs)'s inputs object instead of silently dropping them into the options arg / REST query string. This fixes reasoning models (GLM-4.7-flash, Kimi K2.5/K2.6, GPT-OSS, QwQ) burning the entire output token budget on chain-of-thought with no visible content.

  Both settings-level and per-call usage are supported:

  // Settings-level
  const model = workersai("@cf/zai-org/glm-4.7-flash", {
  	reasoning_effort: "low",
  	chat_template_kwargs: { enable_thinking: false },
  });
  
  // Per-call (overrides settings)
  await generateText({
  	model,
  	prompt,
  	providerOptions: {
  		"workers-ai": { reasoning_effort: "low" },
  	},
  });

  reasoning_effort: null is preserved as-is (explicit "disable reasoning" signal). The two fields are also typed directly on WorkersAIChatSettings.

  Closes #501.