Compare AI models
| Specifications | ||
ParametersThe 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. | — | 9B |
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 | ||
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. | 28.9% | 28.5% |
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. | 12.5% | — |
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% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 51.5% | — |
SimpleQAKnowledge — accuracy on short, fact-seeking questions in the original SimpleQA set | 21.7% | — |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 71.6% | — |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 68% | — |
| Overview | ||
| Company | Z.ai | |
| Release date | Feb 5 2025 | Jul 5 2024 |
| Access | Closed | Open Weight |
| Model details | View model | View model |
Other comparisons
Gemini 2.0 Flash-LitevsClaude Fable 5.2CodeGeeX4-ALL-9BvsClaude Fable 5.2Gemini 2.0 Flash-LitevsGPT-6.2 AstraCodeGeeX4-ALL-9BvsGPT-6.2 AstraGemini 2.0 Flash-LitevsMuse EmberCodeGeeX4-ALL-9BvsMuse EmberGemini 2.0 Flash-LitevsGrok 4.7CodeGeeX4-ALL-9BvsGrok 4.7Gemini 2.0 Flash-LitevsDeepSeek-V4.1-Flash-1017CodeGeeX4-ALL-9BvsDeepSeek-V4.1-Flash-1017Gemini 2.0 Flash-LitevsMistral Large 4CodeGeeX4-ALL-9BvsMistral Large 4Frequently asked questions
Gemini 2.0 Flash-Lite leads CodeGeeX4-ALL-9B on 1 of the 1 benchmark they both report (LiveCodeBench). CodeGeeX4-ALL-9B shipped 215 days before Gemini 2.0 Flash-Lite, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Gemini 2.0 Flash-Lite) vs 128k (CodeGeeX4-ALL-9B). Gemini 2.0 Flash-Lite is closed, while CodeGeeX4-ALL-9B is open weight.
On LiveCodeBench, Gemini 2.0 Flash-Lite leads at 28.9% vs CodeGeeX4-ALL-9B at 28.5%.