ScrapeGraphAI/Scrapegraph-ai · error · ValueError

No pricing tier matches {input_tokens} input tokens

Error message

No pricing tier matches {input_tokens} input tokens

What it means

ValueError from get_model_cost_per_1k_tokens: none of the pricing tiers for the model matched the given input_tokens. Tiers are defined via input_tokens_lte / input_tokens_gt bounds; if the tier table has gaps (e.g. no open-ended top tier), large token counts fall through.

Source

Thrown at scrapegraphai/utils/model_costs.py:179

        raise ValueError("input_tokens must not be negative")

    if model_name in MODEL_COST_TIERS_PER_1K_TOKENS:
        try:
            pricing_tiers = MODEL_COST_TIERS_PER_1K_TOKENS[model_name][service_tier]
        except KeyError as exc:
            raise ValueError(
                f"Unsupported service tier {service_tier!r} for {model_name}"
            ) from exc

        rate_key = "output" if is_completion else "input"
        for pricing in pricing_tiers:
            upper_bound = pricing.get("input_tokens_lte")
            lower_bound = pricing.get("input_tokens_gt")
            if upper_bound is not None and input_tokens <= upper_bound:
                return float(pricing[rate_key])
            if lower_bound is not None and input_tokens > lower_bound:
                return float(pricing[rate_key])
        raise ValueError(f"No pricing tier matches {input_tokens} input tokens")

    costs = (
        MODEL_COST_PER_1K_TOKENS_OUTPUT
        if is_completion
        else MODEL_COST_PER_1K_TOKENS_INPUT
    )
    return costs[model_name]

View on GitHub (pinned to 532dfffbf6)

Solutions

1. Fix the tier definition to include an open-ended top tier (only input_tokens_gt, no lte)
2. Verify the input_tokens value is sane (not multiplied/overflowed)
3. Add a final catch-all tier covering the missing range

Example fix

# before
MODEL_COST_TIERS_PER_1K_TOKENS["m"] = {"standard": [{"input_tokens_lte": 1000000, ...}]}
# after
MODEL_COST_TIERS_PER_1K_TOKENS["m"] = {"standard": [{"input_tokens_lte": 1000000, ...}, {"input_tokens_gt": 1000000, ...}]}

tiers = MODEL_COST_TIERS_PER_1K_TOKENS.get(model, {}).get(service_tier, [])
def matches(t, n):
    return (t.get("input_tokens_lte") is not None and n <= t["input_tokens_lte"]) or (t.get("input_tokens_gt") is not None and n > t["input_tokens_gt"])
assert any(matches(t, input_tokens) for t in tiers) or not tiers

try:
    rate = get_model_cost_per_1k_tokens(model, n)
except ValueError as e:
    rate = float(tiers[-1]["output" if is_completion else "input"])  # top tier fallback

When it happens

Trigger: Passing an input_tokens value outside every defined tier bound for a tiered model — typically a very large count when the top tier lacks an input_tokens_gt-only entry.

Common situations: Hand-edited tier tables missing an unbounded final tier, or token counts exceeding the highest input_tokens_lte.

Related errors

• Unsupported service tier {service_tier!r} for {model_name}
• input_tokens is required for completion costs with tiered pr
• input_tokens must not be negative

AI-assisted analysis of ScrapeGraphAI/Scrapegraph-ai@532dfffbf6 (2026-08-28). Data as JSON: /api/errors/71923f2592908549. Report an issue: GitHub.