{"id":"chitra-tiny","slug":"chitra-tiny","name":"Chitra Tiny","orgId":"udgam-labs","family":"chitra-tiny","releaseDate":null,"availability":"announced","openWeights":true,"license":null,"paramsTotal":null,"paramsSource":null,"contextWindow":null,"knowledgeCutoff":null,"modelCardUrl":"https://udgamlabs.com/blog/chitra-tiny","hfRepo":null,"modalities":["image"],"aliases":["Chitra Tiny","chitra-tiny","Chitra-Tiny","udgamlabs/Chitra-Tiny"],"aliasNote":null,"origin":"tracked","tracked":{"trackedSince":"2026-10-24","announced":"2026-10-12","blurb":"Udgam Labs' 212M-parameter English text-to-image model that makes 256×256 pictures from a sentence. The diffusion transformer was trained from scratch on about 4.7 million public-domain and CC0 images; it uses Flan-T5-Base and Stability AI's sd-vae-ft-mse unchanged. Apache-2.0.","api":"none","openWeights":"yes","params":{"text":"212M (diffusion transformer, 16 layers, width 768)","source":"Udgam Labs model page"},"facts":[{"label":"Method","value":"Latent rectified-flow diffusion transformer with classifier-free guidance; 50 steps by default"},{"label":"Training","value":"Transformer trained from scratch (random weights) on about 4.7 million public-domain and CC0 images with automatically generated captions"},{"label":"Frozen components","value":"google/flan-t5-base text encoder (Apache-2.0) and stabilityai/sd-vae-ft-mse decoder (MIT), used unchanged"},{"label":"Output","value":"256×256 images from English prompts; no built-in safety filter"},{"label":"Weights","value":"Hugging Face (udgamlabs/Chitra-Tiny, about 0.85 GB) and code on GitHub (udgamlabs/chitra-tiny)"},{"label":"License","value":"Apache-2.0 (weights and code)"}],"claims":["Lab claim: GenEval overall 0.219 with the official code, against 0.426 for Stable Diffusion 1.5 in the same harness.","Lab claim: FID 22.1 on 5,000 held-out images like its training data and CLIP score 26.2 (real held-out images: 27.5); Udgam notes both favour the model.","Lab claim: about three seconds per image on an Apple-silicon Mac."],"sources":[{"kind":"lab","title":"Chitra Tiny, a 212M-parameter text-to-image model trained from scratch (Udgam Labs blog)","url":"https://udgamlabs.com/blog/chitra-tiny","date":"2026-10-12"},{"kind":"lab","title":"Chitra Tiny model page (Udgam Labs)","url":"https://udgamlabs.com/models/chitra-tiny","date":"2026-10-12"},{"kind":"repo","title":"udgamlabs/Chitra-Tiny model card (Hugging Face)","url":"https://huggingface.co/udgamlabs/Chitra-Tiny"},{"kind":"repo","title":"udgamlabs/chitra-tiny (GitHub)","url":"https://github.com/udgamlabs/chitra-tiny"}],"firstBenchmark":null},"organization":{"id":"udgam-labs","name":"Udgam Labs","slug":"udgam-labs","country":null,"url":"https://udgamlabs.com","color":"#11110F","mark":"U"},"price":null,"answer":{"word":"Pending benchmarks.","scope":"Announced Oct 12, 2026; no independent results yet.","lead":"As of Nov 3, 2026, Chitra Tiny is pending benchmarks on isitsota: announced Oct 12, 2026, with no independent results on the boards we track yet.","confidence":"low"},"status":"PENDING_BENCHMARKS","verdicts":[]}