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. | 8B | 110B |
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. | 8k | 32k |
| 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. | 36.22% | 52.4% |
HellaSwagReasoning — commonsense completion of short scenarios | 83.03% | 87.5% |
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. | 69.51% | 80.4% |
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. | 58.45% | 85.4% |
MATHReasoning — accuracy on the original competition mathematics test set | — | 49.6% |
ARC-ChallengeReasoning — multiple-choice science questions from the AI2 Reasoning Challenge; unrelated to ARC-AGI | 64.42% | — |
WinoGrandeReasoning — resolving ambiguous pronouns in context | 80.35% | — |
| Overview | ||
| Company | NVIDIA | Qwen |
| Release date | Aug 21 2024 | Apr 25 2024 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
Mistral NeMo Minitron 8BvsClaude Fable 5.2Qwen1.5-110BvsClaude Fable 5.2Mistral NeMo Minitron 8BvsGPT-6.2 AstraQwen1.5-110BvsGPT-6.2 AstraMistral NeMo Minitron 8BvsGemini 4 NeonQwen1.5-110BvsGemini 4 NeonMistral NeMo Minitron 8BvsMuse EmberQwen1.5-110BvsMuse EmberMistral NeMo Minitron 8BvsGrok 4.7Qwen1.5-110BvsGrok 4.7Mistral NeMo Minitron 8BvsDeepSeek-V4.1-Flash-1017Qwen1.5-110BvsDeepSeek-V4.1-Flash-1017Frequently asked questions
Qwen1.5-110B leads Mistral NeMo Minitron 8B on 4 of the 4 benchmarks they both report (HumanEval, HellaSwag, MMLU, GSM8K). Qwen1.5-110B shipped 118 days before Mistral NeMo Minitron 8B, so benchmark comparisons should account for the intervening progress.
Mistral NeMo Minitron 8B has 8B parameters, while Qwen1.5-110B has 110B. Context windows are 8k (Mistral NeMo Minitron 8B) vs 32k (Qwen1.5-110B).
On HumanEval, Qwen1.5-110B leads at 52.4% vs Mistral NeMo Minitron 8B at 36.22%. On HellaSwag, Qwen1.5-110B leads at 87.5% vs Mistral NeMo Minitron 8B at 83.03%. On MMLU, Qwen1.5-110B leads at 80.4% vs Mistral NeMo Minitron 8B at 69.51%. On GSM8K, Qwen1.5-110B leads at 85.4% vs Mistral NeMo Minitron 8B at 58.45%.