We are excited to announce that GLM-5.2 is now available on SCRP-Chat. GLM-5.2 is widely considered to be the first open model competitive with frontier models from OpenAI and Anthropic:

GLM-5.2 Benchmarks Source: https://z.ai/blog/glm-5.2

GLM-5.2 has replaced GLM-5.1 as the model behind SCRP-Assistant.

A few notes about using GLM-5.2 on SCRP-Chat:

  • Due to a bug in vLLM, context length is capped at 300K.
  • Reasoning effort can be adjusted between two levels: max and high, with max being the default. Reasoning can also be completely turned off.
    • To set reasoning effort to high: 
        client.chat.completions.create(
  model="scrp-assistant",
  messages=msgs,
  extra_body={"chat_template_kwargs": {"reasoning_effort": "high"}},
  )
    • To turn off reasoning: 
        client.chat.completions.create(
  model="scrp-assistant",
  messages=msgs,
  extra_body={"chat_template_kwargs": {"enable_thinking": False}},
  )