ChatGLM2-6BOpen Weight

Released

ChatGLM2-6B is an AI model released by Z.ai on Sunday, Jun 25 2023, 103 days after ChatGLM-6B. It is an open-weight model — the trained weights are available to download and run. It is a 6B 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 32k 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 HumanEval, MATH, HellaSwag, MMLU, and GSM8K.

Benchmarks

Coding

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.
9.8%
#73 of 74

Reasoning & science

MATHReasoning — accuracy on the original competition mathematics test set
6.9%
#49 of 50
HellaSwagReasoning — commonsense completion of short scenarios
57%
#16 of 17
MMLUGeneral knowledge — A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better.
45.2%
#67 of 68
GSM8KGrade-school math — Grade-school maths word problems that take a few steps of arithmetic to work through. It separated the models of 2022 and 2023 sharply, then saturated. One caveat on the historical numbers: OpenAI included part of the GSM8K training set in GPT-4's pre-training mix, so GPT-4's score is not a clean few-shot result. Higher is better.
25.9%
#38 of 39

Compare ChatGLM2-6B with

ChatGLM2-6B

About

ChatGLM2-6B was the second generation of Zhipu AI and Tsinghua's bilingual assistant. The team trained a new checkpoint and used multi-query attention to make inference faster. FlashAttention helped extend the sequence length beyond the first generation's 2K tokens.

The initial model supported sequences up to 32K, with dialogue alignment trained at 8K. The GLM team's later family report measured 45.2% on MMLU and 25.9% on GSM8K for this checkpoint.

Frequently asked questions

ChatGLM2-6B was released by Z.ai on Sunday, Jun 25 2023.

All Z.ai releases

24 tracked