Compare AI models
| Specifications | ||
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. | 1M | 128k |
| Benchmarks | ||
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. | 74.3% | 75.7% |
MATHReasoning — accuracy on the original competition mathematics test set | 54.9% | 57.9% |
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. | 78.9% | 81.9% |
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. | 86.2% | 90.9% |
SWE-Bench VerifiedCoding — Real coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 9.6% | — |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 75.2% |
HellaSwagReasoning — commonsense completion of short scenarios | 86.5% | — |
MGSMMultilingual math — solving grade-school word problems translated into multiple languages | 82.6% | — |
MathVistaMultimodal reasoning — mathematical reasoning about visual information | 58.4% | — |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 56.1% | — |
| Overview | ||
| Company | Z.ai | |
| Release date | May 14 2024 | Jun 5 2024 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Gemini 1.5 FlashvsClaude Fable 5.2GLM-4-AirvsClaude Fable 5.2Gemini 1.5 FlashvsGPT-6.2 AstraGLM-4-AirvsGPT-6.2 AstraGemini 1.5 FlashvsMuse EmberGLM-4-AirvsMuse EmberGemini 1.5 FlashvsGrok 4.7GLM-4-AirvsGrok 4.7Gemini 1.5 FlashvsDeepSeek-V4.1-Flash-1017GLM-4-AirvsDeepSeek-V4.1-Flash-1017Gemini 1.5 FlashvsMistral Large 4GLM-4-AirvsMistral Large 4Frequently asked questions
GLM-4-Air leads Gemini 1.5 Flash on 4 of the 4 benchmarks they both report (HumanEval, MATH, MMLU, GSM8K). Gemini 1.5 Flash shipped 22 days before GLM-4-Air, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Gemini 1.5 Flash) vs 128k (GLM-4-Air).
On HumanEval, GLM-4-Air leads at 75.7% vs Gemini 1.5 Flash at 74.3%. On MATH, GLM-4-Air leads at 57.9% vs Gemini 1.5 Flash at 54.9%. On MMLU, GLM-4-Air leads at 81.9% vs Gemini 1.5 Flash at 78.9%. On GSM8K, GLM-4-Air leads at 90.9% vs Gemini 1.5 Flash at 86.2%.