CodeGeeX4-ALL-9BOpen Weight

Released

CodeGeeX4-ALL-9B is an AI model released by Z.ai on Friday, Jul 5 2024, 30 days after GLM-4-Air. It is an open-weight model — the trained weights are available to download and run. It is a 9B parameterThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. model with a 128k token context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider.. Benchmark results (shown below) cover LiveCodeBench, and HumanEval.

Benchmarks

LiveCodeBenchCompetitive coding — Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better.
28.5%
#69 of 70
HumanEvalFunction synthesis — 164 small Python problems: the AI is given a function's description and has to write the working function. This was the coding benchmark of the GPT-3.5 and GPT-4 era, before the field moved to fixing real bugs in real repositories. Higher is better.
82.3%
#33 of 74

Compare CodeGeeX4-ALL-9B with

CodeGeeX4-ALL-9B

About

CodeGeeX4-ALL-9B was released in July 2024 after continued training from GLM-4-9B. The CodeGeeX team designed the model to work across the editor tasks of code completion, generation and questions about a repository.

The model supported function calling and could work with a 128K context. Its published evaluation reported 82.3% on HumanEval, with a separate evaluation for fill-in-the-middle tasks that supply code on both sides of a cursor.

Frequently asked questions

CodeGeeX4-ALL-9B was released by Z.ai on Friday, Jul 5 2024.

All Z.ai releases

24 tracked