xtekky/gpt4free · error · ValueError

Thinking mode must be an integer between 0 and 4

Error message

Thinking mode must be an integer between 0 and 4

What it means

ValueError from Gemini._resolve_model: a thinking level was resolved (via @think= suffix or think_override) but it is not an integer in the inclusive range 0-4. Gemini's thinking-mode knob only accepts five discrete levels, so anything outside that band — or a non-int type from think_override — is rejected before any request is made.

Source

Thrown at g4f/Provider/needs_auth/Gemini.py:253

            line, buffer = buffer.split(b"\n", 1)
            yield line.decode("utf-8", errors="replace")
    if buffer:
        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")

View on GitHub (pinned to 973504e177)

Solutions

1. Clamp the think level to 0-4, e.g. "@think=3".
2. Pass think_override as a plain int, not a string or bool.
3. Validate user-supplied think values before forwarding them to the provider.

Example fix

# before
model = "gemini-2.5-flash@think=9"  # ValueError: Thinking mode must be an integer between 0 and 4

# after
model = "gemini-2.5-flash@think=4"

def clamp_think_level(value):
    n = int(value)
    if not 0 <= n <= 4:
        raise ValueError(f"think level {n} out of range 0-4")
    return n

safe_level = clamp_think_level(user_input)

def is_valid_think_override(value) -> bool:
    return type(value) is int and 0 <= value <= 4

try:
    resp = await client.chat.completions.create(model=m, provider=Gemini, messages=msgs, think_override=lvl)
except ValueError as e:
    if "between 0 and 4" in str(e):
        lvl = max(0, min(4, lvl))
        resp = await client.chat.completions.create(model=m, provider=Gemini, messages=msgs, think_override=lvl)

When it happens

Trigger: Passing @think=7 or @think=-1 in the model name, or a think_override that is a bool/float/string. Note the is-instance check also catches values like True (bool is an int subclass but the isinstance gate is on int, so bools and other types fail).

Common situations: Guessing a 0-10 scale from other providers; passing think_override as a string "2"; copy-pasted configs from tools using a different level range.

Related errors

• Invalid thinking mode: {think_value!r}
• Unknown Gemini model: {model}. Supported models: {', '.join(
• No refresh token found in GCP_SERVICE_ACCOUNT.
• No project information found in API response.
• No Yupp accounts configured. Set YUPP_API_KEY environment va

AI-assisted analysis of xtekky/gpt4free@973504e177 (2026-08-14). Data as JSON: /api/errors/e49516e0e97f28d7. Report an issue: GitHub.