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. | 8k | — |
| API pricingUSD per 1M tokens · lower wins | ||
Cheapest inputLowest input rate across every provider on OpenRouter, the lab included. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $30.00Azure | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $60.00Azure | — |
| $9,000Azure | — | |
| 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. | 67% | 85.7% |
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. | 86.4% | 86.2% |
MATHReasoning — accuracy on the original competition mathematics test set | — | 73% |
DROPReasoning — three-shot token F1 on questions requiring reasoning over paragraphs | 80.9 | — |
ARC-ChallengeReasoning — multiple-choice science questions from the AI2 Reasoning Challenge; unrelated to ARC-AGI | 96.3% | — |
HellaSwagReasoning — commonsense completion of short scenarios | 95.3% | — |
WinoGrandeReasoning — resolving ambiguous pronouns in context | 87.5% | — |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | — | 72% |
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. | 92% | — |
MathVistaMultimodal reasoning — mathematical reasoning about visual information | — | 68.1% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | — | 63.2% |
| Overview | ||
| Company | OpenAI | SpaceXSI |
| Release date | Mar 14 2023 | Aug 13 2024 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
GPT-4vsClaude Fable 5.2Grok‑2 minivsClaude Fable 5.2GPT-4vsGemini 4 HydrogenGrok‑2 minivsGemini 4 HydrogenGPT-4vsMuse EmberGrok‑2 minivsMuse EmberGPT-4vsDeepSeek-V4.1-Flash-1017Grok‑2 minivsDeepSeek-V4.1-Flash-1017GPT-4vsMistral Large 4Grok‑2 minivsMistral Large 4GPT-4vsKimi K3.1 ProGrok‑2 minivsKimi K3.1 ProFrequently asked questions
GPT-4 and Grok‑2 mini are evenly matched across the 2 benchmarks they both report (HumanEval, MMLU). GPT-4 shipped 518 days before Grok‑2 mini, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On HumanEval, Grok‑2 mini leads at 85.7% vs GPT-4 at 67%. On MMLU, GPT-4 leads at 86.4% vs Grok‑2 mini at 86.2%.