“Gemini” is a family, not one score. On the 2026-08-02 Artificial Analysis snapshot:
- Video — Gemini Omni Flash leads text-to-video (1245) and sits near the top of image-to-video. Restored pairing: Gemini Omni Flash vs MiniMax H3
- Image — Nano Banana 2 / Gemini 3.1 Flash Image variants sit in the image and image-editing boards
- TTS — Gemini 3.1 Flash TTS is in the text-to-speech top five
- LLMs — Anthropic Opus 5 and OpenAI GPT-5.6 Sol currently hold the top LLM ranks; Gemini’s chat models compete in product surfaces even when they’re not #1 on this particular intelligence index
When to reach for it
Long context, multimodal dumps (docs + video + logs), and Google-ecosystem workflows. For pure coding-agent wins, compare Opus 5 and GPT-5.6 Sol on the LLM board first — see /posts/claude-vs-gpt5.
What we do not claim
We ingest Artificial Analysis, not Google’s marketing pages. If Gemini’s chat rank moves on AA’s intelligence index, /leaderboard updates on the next snapshot. This post is a map of where Gemini shows up across modalities, not a verdict that Gemini is “best overall.”
Source: /leaderboard. Methodology: /methodology.