Compare models
Pick up to 4models. Specs render side-by-side. Share the URL — it's stateless.
Comparing agents instead? Switch to agent compare →
Meta | |||
|---|---|---|---|
| Vendor | Meta | ||
| Family | Gemini | Llama | Gemini |
| Release date | 2025-03-25 | 2025-04-05 | 2026-05-20 |
| Context window | 2,000,000 tokens | 10,000,000 tokens | 2,000,000 tokens |
| Parameters | — | 109B | — |
| Modality | text, vision, audio, video | text, vision | text, vision, audio, video |
| License | proprietary | Llama 4 Community License | proprietary |
| Source | proprietary | open weights | proprietary |
| Description | Predecessor to Gemini 3 Pro. | — | Google DeepMind's next-gen Gemini — positioned as "frontier intelligence with action". Built for complex agentic workflows. Announced at Google I/O 2026. |
| Links | |||
| Benchmarks | |||
| MMLU-Pro | 84.1% | — | 88.9% |
| GPQA-D | 84.0% | — | 91.2% |
| AIME-25 | — | — | 96.1% |
| MMMU | 72.4% | — | 86.0% |