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. | 2B/7B | — |
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 | 200k |
| These models have no shared benchmark scores. | ||
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. | 32.3% | — |
MATHReasoning — accuracy on the original competition mathematics test set | 24.3% | — |
ARC-ChallengeReasoning — multiple-choice science questions from the AI2 Reasoning Challenge; unrelated to ARC-AGI | 53.2% | — |
HellaSwagReasoning — commonsense completion of short scenarios | 81.2% | — |
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. | 64.3% | — |
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. | 46.4% | — |
| Overview | ||
| Company | OpenAI | |
| Release date | Feb 21 2024 | May 16 2025 |
| Access | Open Weight | Closed |
| Model details | View model | View model |
Other comparisons
Gemma 1vsClaude Fable 5.2codex-mini-latestvsClaude Fable 5.2Gemma 1vsMuse Embercodex-mini-latestvsMuse EmberGemma 1vsGrok 4.7codex-mini-latestvsGrok 4.7Gemma 1vsDeepSeek-V4.1-Flash-1017codex-mini-latestvsDeepSeek-V4.1-Flash-1017Gemma 1vsMistral Large 4codex-mini-latestvsMistral Large 4Gemma 1vsKimi K3.1 Procodex-mini-latestvsKimi K3.1 ProFrequently asked questions
Gemma 1 and codex-mini-latest don't publish scores on any of the same benchmarks, so there's no direct head-to-head comparison. Gemma 1 shipped 450 days before codex-mini-latest, so benchmark comparisons should account for the intervening progress.
Context windows are 8k (Gemma 1) vs 200k (codex-mini-latest). Gemma 1 is open weight, while codex-mini-latest is closed.
Direct benchmark comparisons are unavailable — Gemma 1 and codex-mini-latest don't publish scores on any of the same benchmarks.