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. | — | 7B/14B |
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. | 2M | 1M |
| 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. | 36% | 38.6% |
MATHReasoning — accuracy on the original competition mathematics test set | 91.8% | 79.5% |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 79.1% | 63.3% |
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. | — | 88.4% |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 84.3% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 64.7% | — |
SimpleQAKnowledge — accuracy on short, fact-seeking questions in the original SimpleQA set | 44.3% | — |
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. | — | 94.8% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 72.7% | — |
| Overview | ||
| Company | Qwen | |
| Release date | Feb 5 2025 | Jan 27 2025 |
| Access | Closed | Open Weight |
| Model details | View model | View model |
Other comparisons
Gemini 2.0 ProvsClaude Fable 5.2Qwen2.5-1MvsClaude Fable 5.2Gemini 2.0 ProvsGPT-6.2 AstraQwen2.5-1MvsGPT-6.2 AstraGemini 2.0 ProvsMuse EmberQwen2.5-1MvsMuse EmberGemini 2.0 ProvsGrok 4.7Qwen2.5-1MvsGrok 4.7Gemini 2.0 ProvsDeepSeek-V4.1-Flash-1017Qwen2.5-1MvsDeepSeek-V4.1-Flash-1017Gemini 2.0 ProvsMistral Large 4Qwen2.5-1MvsMistral Large 4Frequently asked questions
Gemini 2.0 Pro leads Qwen2.5-1M on 2 of the 3 benchmarks they both report (LiveCodeBench, MATH, MMLU-Pro). Qwen2.5-1M shipped 9 days before Gemini 2.0 Pro, so benchmark comparisons should account for the intervening progress.
Context windows are 2M (Gemini 2.0 Pro) vs 1M (Qwen2.5-1M). Gemini 2.0 Pro is closed, while Qwen2.5-1M is open weight.
On LiveCodeBench, Qwen2.5-1M leads at 38.6% vs Gemini 2.0 Pro at 36%. On MATH, Gemini 2.0 Pro leads at 91.8% vs Qwen2.5-1M at 79.5%. On MMLU-Pro, Gemini 2.0 Pro leads at 79.1% vs Qwen2.5-1M at 63.3%.