xtekky/gpt4free · error · ValueError

Unknown Gemini model: {model}. Supported models: {', '.join(

Error message

Unknown Gemini model: {model}. Supported models: {', '.join(models)}

What it means

ValueError from Gemini._resolve_model: after alias expansion (MODEL_ALIASES) the model name is not in the provider's supported models set. The message lists every supported model, so it doubles as live documentation. It fires before any network call — a pure input-validation failure.

Source

Thrown at g4f/Provider/needs_auth/Gemini.py:256

        yield buffer.decode("utf-8", errors="replace")


def _resolve_model(model: str, think_override: int = None) -> tuple[str, bool]:
    requested_model = model
    think_mode = think_override
    if "@think=" in model:
        model, think_value = model.rsplit("@think=", 1)
        requested_model = model
        try:
            think_mode = int(think_value)
        except ValueError as exc:
            raise ValueError(f"Invalid thinking mode: {think_value!r}") from exc
    if think_mode is not None:
        if not isinstance(think_mode, int) or not 0 <= think_mode <= 4:
            raise ValueError("Thinking mode must be an integer between 0 and 4")
    model = MODEL_ALIASES.get(model, model)
    if model not in models:
        raise ValueError(
            f"Unknown Gemini model: {model}. " f"Supported models: {', '.join(models)}"
        )
    expanded_thinking = (
        requested_model in EXPANDED_MODEL_ALIASES
        if think_mode is None
        else think_mode <= 2
    )
    return model, expanded_thinking


def _normalize_messages(messages: Messages | None) -> Messages:
    if messages is None:
        return []
    if not isinstance(messages, list):
        raise TypeError("messages must be a list")
    return messages

View on GitHub (pinned to 973504e177)

Solutions

1. Pick a model from the list embedded in the error message and use that exact string.
2. Update g4f to the latest version so newly released Gemini models are recognized.
3. Check MODEL_ALIASES in the provider source for accepted shorthand names before inventing your own.
4. If the model must be user-configurable, validate against the provider's models list at startup.

Example fix

# before
resp = await client.chat.completions.create(model="gemini-2.5-flsh", provider=Gemini, messages=msgs)
# ValueError: Unknown Gemini model: gemini-2.5-flsh. Supported models: ...

# after
resp = await client.chat.completions.create(model="gemini-2.5-flash", provider=Gemini, messages=msgs)

from g4f.Provider.needs_auth.Gemini import Gemini
from g4f.models import models as known_models  # adjust to your g4f version

def gemini_model_supported(model: str) -> bool:
    base = model.rsplit("@think=", 1)[0]
    aliases = getattr(Gemini, "MODEL_ALIASES", {})
    resolved = aliases.get(base, base)
    return resolved in Gemini.supported_models if hasattr(Gemini, "supported_models") else True

def is_known_gemini_model(model: str, supported: set[str]) -> bool:
    base = model.rsplit("@think=", 1)[0]
    return base in supported

try:
    resp = await client.chat.completions.create(model=m, provider=Gemini, messages=msgs)
except ValueError as e:
    if "Unknown Gemini model" in str(e):
        supported = [name for name in str(e).split("Supported models: ")[-1].split(", ")]
        m = pick_closest(m, supported)

When it happens

Trigger: Passing a Gemini model string that is neither a known alias nor a supported model — typos ("gemini-2.5-flsh"), retired names, or names from other providers. Also triggers when an @think= suffix was stripped correctly but the base name is wrong.

Common situations: Model deprecated/renamed after a Gemini API update while g4f pins the old list; typo'd model names from config files; assuming any Google model id works.

Related errors

• Invalid thinking mode: {think_value!r}
• Thinking mode must be an integer between 0 and 4
• Model '{model}' is not supported by {cls.__name__}. Supporte
• Provider '{item}' not found
• Label must be provided

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/0fe5ebe95818e474. Report an issue: GitHub.