Two Google models are available via the API but missing from the models.dev database:
gemini-2.5-flash-tts — Gemini 2.5 Flash text-to-speech model. Produces audio output (audio/ MIME types) rather than text tokens.gemini-3-pro-image-preview — Gemini 3 Pro image generation preview model.
Both are present in Google's inference routing (confirmed via the generativelanguage API) but have no corresponding .toml file under models/google/.
For reference, the equivalent files that do exist for similar models:
models/google/gemini-2.5-flash-image.tomlmodels/google/gemini-3.1-flash-image-preview.toml
langchain-google-genai auto-generates its _profiles.py from this repo, so adding these entries would allow the library to correctly populate capability flags (audio_outputs, image_outputs, etc.) for these models.
Two Google models are available via the API but missing from the models.dev database:
gemini-2.5-flash-tts— Gemini 2.5 Flash text-to-speech model. Produces audio output (audio/MIME types) rather than text tokens.gemini-3-pro-image-preview— Gemini 3 Pro image generation preview model.Both are present in Google's inference routing (confirmed via the
generativelanguageAPI) but have no corresponding.tomlfile undermodels/google/.For reference, the equivalent files that do exist for similar models:
models/google/gemini-2.5-flash-image.tomlmodels/google/gemini-3.1-flash-image-preview.tomllangchain-google-genaiauto-generates its_profiles.pyfrom this repo, so adding these entries would allow the library to correctly populate capability flags (audio_outputs,image_outputs, etc.) for these models.