{
 "slug": "Flamingo",
 "title": "Flamingo",
 "type": "model",
 "short_desc": "DeepMind's April 2022 visual language model that pioneered few-shot multimodal prompting by bridging frozen vision and language models.",
 "categories": [
  "Google models",
  "Multimodal models",
  "Models",
  "2022 model releases"
 ],
 "infobox": {
  "Developer": "DeepMind",
  "Announced": "April 2022",
  "Architecture": "Frozen Chinchilla-family Transformer language model bridged to a frozen NFNet vision encoder via a Perceiver Resampler and gated cross-attention layers",
  "Parameters": "3B, 9B, and 80B variants (largest built on the 70B Chinchilla)",
  "Context window": "Interleaved image-text sequences; token limit undisclosed",
  "Post-training": "None for the flagship; trained on interleaved web data, image-text and video-text pairs",
  "License / access": "Proprietary; research model, never publicly released",
  "Paper/report": "arXiv:2204.14198 (NeurIPS 2022)"
 },
 "infobox_links": {
  "Developer": [
   "Google_DeepMind"
  ],
  "Architecture": [
   "Chinchilla",
   "Transformer_(architecture)"
  ],
  "Parameters": [
   "Chinchilla"
  ]
 },
 "aliases": [
  "Flamingo (model)"
 ],
 "family": "",
 "org": "",
 "registry_status": "stub",
 "words": 368,
 "references": 3,
 "outbound": [
  "CLIP",
  "Chinchilla",
  "GPT-3",
  "Gemini_(model_family)",
  "Google_DeepMind",
  "Hugging_Face",
  "IDEFICS",
  "Large_language_model",
  "Moondream_(model_family)",
  "PaLM",
  "PaliGemma",
  "Transformer_(architecture)"
 ],
 "inbound": [
  "BLIP",
  "CLIP",
  "IDEFICS",
  "Moondream_(model_family)",
  "Vision_Transformer"
 ],
 "url": "wiki/Flamingo.html",
 "built": "2026-07-24"
}