LLM RankHUB
LLM price intelligence · BYOK gateway

The same model,
for less,
more reliably.

Compare 780+ channels, rank by real production token volume, and see free quotas and where every token goes.

600+models
780+channels
130+free channels
Dailyreal usage refresh
Open llmrankhub.com → View GitHub
Every price links to its source · conservative free-tier checks · prompts are never logged
Token MeterBYOK gatewayAES-256
Token Meter: spend, token breakdown, model ranking and cache hit rate
63.2%Cache hit rate · live metering
Why LLM Rank Hub

Not another leaderboard.
A procurement and routing desk.

Cheaper, more reliable, or still free — all in one table. Compare first, then verify with real usage.

Real usage

Ranked by real production token volume — today, this week, this month. Not a vote.

Cross-channel pricing

Official, measured, peak/off-peak and cache prices across 780+ channels, each linked to its source.

Free tier

Only models with a free channel. Open one to see free channels only, with daily / per-minute quotas.

BYOK Token Meter

Host your own key and see spend, cache hit rate, latency and model mix.

The difference

Connect your key.
See where every token goes.

Most price sites stop at the list price. LLM Rank Hub goes one step further: real usage, real cost structure.

  • Key custody: AES-256-CTR + HMAC. The master key lives outside the database; deleting a key wipes its ciphertext.
  • One gateway, two protocols: /v1/chat/completions and /v1/messages, with an llmrh_… token.
  • Real metering: input / output / cached / reasoning tokens, cost, cache hit rate, TPS, TTFT and per-request events.
  • No content logging: prompts and replies are relayed in memory only; never written to disk or logs.
Open Token Meter →
# OpenAI SDK · Anthropic SDK works too from openai import OpenAI client = OpenAI( base_url="https://llmrankhub.com/v1", api_key="llmrh_...", )
Token Meter: spend, token breakdown, model ranking and recent requests
Sample account data · actual values depend on your key
Two boards

Start with real usage.
Then see where it is still free.

One model, two views: real production token volume, and the free channels and quotas still available.

Real usage leaderboard
REAL USAGE

Real usage board

Today / this week / this month, with CN / intl and open / proprietary filters; expand for per-channel price, uptime and task cost.

Free tier: models with a free channel
FREE TIER

Free tier board

Only models with at least one free channel. Open one to see free channels only, with daily / per-minute quotas.

Coverage

The models you know.
One procurement table.

Integrated channels and model vendors
Selected channels and model vendors · no particular order · full list on the live boards
Method

Numbers, not guesses.

Prices you can check

Every price links to the vendor's pricing page or a verified source. No source, no number. The cheapest row per vendor and product line wins.

Real usage

The usage board ranks by production token volume — today, this week, this month. Not votes, not visits.

Conservative free tiers

A zero price plus an explicit free label or a verified source. Missing prices are never treated as free.

Compare first. Then decide.

Open llmrankhub.com to find the best channel for your model, then connect a key to verify the real cost.

Open llmrankhub.com →