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. | 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. | 30% | 74.8% |
MATHReasoning — accuracy on the original competition mathematics test set | 77.9% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 51% | — |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 67.3% | — |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 62.3% | — |
| Overview | ||
| Company | SpaceXSI | |
| Release date | Sep 24 2024 | Feb 17 2025 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Gemini 1.5 Flash-002vsClaude Fable 5.2Grok 3 minivsClaude Fable 5.2Gemini 1.5 Flash-002vsGPT-6.1 TerraGrok 3 minivsGPT-6.1 TerraGemini 1.5 Flash-002vsMuse EmberGrok 3 minivsMuse EmberGemini 1.5 Flash-002vsDeepSeek-V4.1-Flash-1017Grok 3 minivsDeepSeek-V4.1-Flash-1017Gemini 1.5 Flash-002vsMistral Large 4Grok 3 minivsMistral Large 4Gemini 1.5 Flash-002vsKimi K3.1 ProGrok 3 minivsKimi K3.1 ProFrequently asked questions
Grok 3 mini leads Gemini 1.5 Flash-002 on 1 of the 1 benchmark they both report (LiveCodeBench). Gemini 1.5 Flash-002 shipped 146 days before Grok 3 mini, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Gemini 1.5 Flash-002) vs 128k (Grok 3 mini).
On LiveCodeBench, Grok 3 mini leads at 74.8% vs Gemini 1.5 Flash-002 at 30%.