compare_gpu_prices
Compare tracked GPU rental prices for one model (e.g. 'H100', 'A100 SXM', 'RTX 4090', 'L40S'), cheapest first. If a result looks wrong, give the user its report_issue_url.
Same data as https://thelanternworks.com/api/v1/gpu-prices.json. Sorted by hourly_usd ascending; each result has source_url and checked_at. 'Cheapest' means cheapest among tracked providers, not the whole market. limit 1-25. Unknown model returns not_found listing tracked models. H100/A100: cheapest_by_form_factor gives one pick per form_factor (tier_class, interruptible, in_stock); same rule as the site. 'NVLink' rows are form 'not stated': NVLink is the bridge, and the vendor doesn't say SXM or PCIe. Each result has tier_class: on-demand | community | marketplace | interruptible (spot/preemptible) | reserved | dedicated-monthly. Reserved (reservation/commitment 'starting from') rates are left out unless include_reserved=true, and even then they are never reported as cheapest. DEFAULT on_demand_only=true: results[0] is the cheapest on-demand rate. Pass on_demand_only=false to include spot/interruptible, community and marketplace capacity (each such row is labelled by tier_class and interruptible); the summary then names the cheapest on-demand rate when the overall cheapest isn't. If a model has no on-demand rate, or every on-demand rate needs more GPUs than requested, the default falls back to all rates and says so in on_demand_fallback. gpus (optional, 1-1024; default 1, stated in assumptions): the GPU count you need. A rate ranks only when its min_gpus (smallest allocation sold) is <= gpus; rates sold only as larger allocations (e.g. 8-GPU nodes) are NOT ranked and are listed in excluded_min_gpus (vendor, gpu, tier, min_gpus, gpu_count, hourly_usd, rate_id, source_url, checked_at, reason, min_bill_usd_per_hour, min_bill_basis); min_gpus_unmet counts them (also in brief). Pass gpus=8 to compare 8-GPU nodes. If every rate needs more GPUs, the error says which gpus value ranks them. Each result has min_bill_usd_per_hour = min_gpus x hourly_usd (min_bill_basis: hourly_rate, per_second_rate or monthly_equivalent) and effective_at, when the vendor started charging this price (effective_at_basis vendor_stated, or first_observed = first daily snapshot showing it after a different price, an upper bound; null when not datable), separate from last_verified. Neither affects ranking. Full results also include batch_due_ledger: each scheduled build (gpu-refresh, lium-collect, scout-import, scout-wider) has expected_at, last_expected_at, and the member_rate_ids it covers, derived only from the workflow schedules. Brief mode omits the ledger. A failed or out-of-stock probe keeps failure_reason (sold_out, region, quota, or other) and failure_reason_detail; both stay on a lapsed claim next to last_known_in_stock. Null when the source does not say; never guessed. Display only. A row with unknown stock, and each stale_prices entry, has next_check_at (next scheduled check, UTC), will_backfill (true: a scheduled check covers it; false: scope limit, nothing will re-check it; null: not derivable), backfill_eta and next_check_basis (the schedule it came from). include_history=true adds each result's price_history (daily-snapshot change points) and history.recent_changes (price changes for this model among tracked providers, last 30 days) from /api/v1/gpu-prices/history.json. Each result has data_age_hours (hours since its price was last verified), degraded (> 48 h) and stale (> 7 days). A degraded row is shown and flagged but never ranked first while a fresh row exists; if none is fresh, stale_fallback=true and the summary says so. By default a rate whose price is older than 168 h is never ranked or named cheapest: it is listed in stale_prices ('stale price (N days old)'); if every rate is stale, results is empty and the summary says no current cheapest. max_data_age_hours (optional) replaces the 168 h default and drops older rates; excluded_too_old counts them. region (optional): a country/area code ('US', 'EU') matches every published zone in it; a zone code ('US-NE-1', 'us-east-1') matches exactly, any case. Rates whose unexpired published regions exclude it are dropped; rates with no region evidence stay, labelled unresolved and ranked after the rest. Each result has eligibility {eligible, applied, unmet, unknown} from the shared T47 contract; constraints.eligible_rate_ids is the same set estimate_job_cost and price watches use. Unknown stock is not confirmed availability. Marketplace median rows (price_status=marketplace) also have marketplace_query (the literal filter used for the median, plus structured filters; null when unknown) next to offer_count and lowest_offer_usd, and composition_flag (composition, price, mixed, unknown, or null) from successive median, minimum and listing count only, not a host intersection. A composition, mixed or unknown flag is not a price move. Missing previous min or n stays unknown, never invented. Stale prices stay flagged as stale. model_identity (T61): sourced nominal_memory and usable_memory (vendor-allocatable) for the queried model, each with source_url; unknown stays 'unknown'. min_usable_memory_gb (optional) keeps only rows whose SOURCED usable memory meets it; rows with unknown usable memory are counted in usable_memory_filter.unknown and never matched on nominal memory or vram_gb. availability_rate_7d (when present): window_days is the planned window (7, not the observed span); observed_hours is the actual coverage from first_check_at to last_check_at (or window_end). coverage_note is set only when observed_hours is shorter than the planned window (e.g. "Covers 24.1h of the planned 7-day window."); otherwise null. availability_window_summary (file-level, also on this output): no_window_count is the number of ranked rates with no 7-day stock window, split by no_stock_source, read_no_usable_signal, and signal_no_window_yet; earliest_first_check_at is when current windows began. content_build_id is a sha256 prefix of canonicalized rate content (timestamps and ids excluded); generation_id is unchanged.
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"gpu": "<gpu>"
},
"name": "compare_gpu_prices"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| gpu | Yes | ||
| gpus | No | ||
| mode | No | brief | |
| limit | No | ||
| region | No | ||
| on_demand_only | No | ||
| include_history | No | ||
| include_reserved | No | ||
| max_data_age_hours | No | ||
| min_usable_memory_gb | No |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| ok | Yes | ||
| gpus | No | ||
| mode | No | ||
| error | No | ||
| results | No | ||
| assumptions | No | ||
| stale_prices | No | ||
| min_gpus_unmet | No | ||
| batch_due_ledger | No | ||
| excluded_min_gpus | No | ||
| referral_disclosure | No | One-line referral disclosure, present whenever the result names a provider; each provider-naming item then has tracked_url (tagged /go/ redirect). |