GLM-4.7-Flash (MoE): zai-org/GLM-4.7-Flash

GLM-4.7-Flash is a mixture of experts (MoE) model from the GLM family with MLA (Multi-head Latent Attention) architecture.

HTTP API

Start the server:

mistralrs serve --isq 4 -p 1234 -m zai-org/GLM-4.7-Flash

Send requests using an OpenAI-compatible client:

import openai

client = openai.Client(base_url="http://localhost:1234/v1", api_key="foobar")

messages = []
prompt = input("Enter system prompt >>> ")
if len(prompt) > 0:
    messages.append({"role": "system", "content": prompt})


while True:
    prompt = input(">>> ")
    messages.append({"role": "user", "content": prompt})
    completion = client.chat.completions.create(
        model="default",
        messages=messages,
        max_tokens=256,
        frequency_penalty=1.0,
        top_p=0.1,
        temperature=0,
    )
    resp = completion.choices[0].message.content
    print(resp)
    messages.append({"role": "assistant", "content": resp})

Python SDK

from mistralrs import Runner, Which, ChatCompletionRequest, Architecture

runner = Runner(
    which=Which.Plain(
        model_id="zai-org/GLM-4.7-Flash",
        arch=Architecture.GLM4MoeLite,
    ),
)

res = runner.send_chat_completion_request(
    ChatCompletionRequest(
        model="default",
        messages=[
            {"role": "user", "content": "Tell me a story about the Rust type system."}
        ],
        max_tokens=256,
        presence_penalty=1.0,
        top_p=0.1,
        temperature=0.1,
    )
)
print(res.choices[0].message.content)
print(res.usage)

Rust SDK

You can find this example here.

use anyhow::Result;
use mistralrs::{IsqType, TextMessageRole, TextMessages, TextModelBuilder};

#[tokio::main]
async fn main() -> Result<()> {
    let model = TextModelBuilder::new("zai-org/GLM-4.7-Flash")
        .with_isq(IsqType::Q4K)
        .with_logging()
        .build()
        .await?;

    let messages = TextMessages::new()
        .add_message(
            TextMessageRole::System,
            "You are an AI agent with a specialty in programming.",
        )
        .add_message(
            TextMessageRole::User,
            "Hello! How are you? Please write generic binary search function in Rust.",
        );

    let response = model.send_chat_request(messages).await?;

    println!("{}", response.choices[0].message.content.as_ref().unwrap());
    dbg!(
        response.usage.avg_prompt_tok_per_sec,
        response.usage.avg_compl_tok_per_sec
    );

    Ok(())
}