COMING SOON — EARLY ACCESS

The model you shipped on changed last month.
Nobody sent you a changelog.

DogKitchen Pro watches the models your product actually runs on — price moves, benchmark shifts, and cheaper equivalents as they appear — and tells you the week it happens instead of the quarter you notice the bill. Same public, citable sources behind AI Pulse, with every snapshot dated and kept.

Public, citable sources
|
API-first
|
CSV & JSON export
|
Every snapshot dated

Real numbers, not a mockup

Everything below is the live snapshot powering AI Pulse right now — same figures, same sources, checkable against LMArena and OpenRouter. Pro adds the history, the alerts, and the API on top.

🎯

Code-Gen Benchmark

💰

Token Price Watch

Cost Efficiency Index

⚖️

Open vs. Closed Parity

best open model as a share of the best closed model's ELO

Six things worth paying for

The free tier tells you where things stand today. Pro tells you when that changes, what it changed from, and what it costs you — the part that's tedious to build and easy to forget to check.

🔔

Change Alerts

Put the models you actually ship on into a watchlist. When input or output pricing moves, when an Arena ranking shifts, or when a model you depend on is beaten on the benchmark you care about, you get a Slack message or a webhook. The email your provider doesn't send.

Price & ELO Slack / Webhook Per-model watchlist
🗄️

The Full Dated Archive

Every snapshot we've ever taken, kept and queryable — rankings, benchmark scores, and per-provider pricing on the date they were true. Answer "what did this cost in March?" or "when did that model overtake ours?" without trusting anyone's memory. The part nobody can backfill.

Point-in-time queries Since July 2026 Growing daily
💰

Cost Modeling for Your Volume

Enter your monthly input and output token volume and see the real monthly bill on every tracked provider, side by side — including the ones you dismissed a year ago that are now cheaper than what you're on. The spreadsheet you keep meaning to rebuild.

Your token volume All tracked providers Monthly projection
🔀

Migration Analysis

Considering a switch? See the quality delta and the cost delta together — Arena ELO, per-benchmark scores, and what your actual volume would cost on each side. Know whether you're trading three points of HumanEval for sixty percent off, or the reverse. Decide with both numbers.

Quality delta Cost delta Side by side
🔌

API Access

Everything above, programmatically. Pull current rankings, pricing, and both original indices into your own dashboards, CI checks, or budget models — or gate a deploy on a benchmark threshold. A REST endpoint, not a PDF.

REST + JSON 10K calls/mo Historical endpoints
📤

Export & Own Your Copy

Take the whole archive with you — CSV or JSON on any tier, Parquet on Enterprise. Drop it into your warehouse, your notebook, or the board deck where someone will ask where the number came from. Every row carries its source and its date.

CSV / JSON Parquet Sourced & dated

The Intelligence Pipeline

Scheduled fetches every six hours, normalized into one schema, diffed against the previous snapshot, then archived under the date they were true.

Sources
🏆 LMArena
🤗 HuggingFace
💱 OpenRouter
🐙 GitHub
📰 RSS feeds
🌐 Currents API
⚙️ Ingestion &
Normalization
🔬 Index Computation &
Quality Scoring
🚨 Snapshot Diffing &
Change Detection
Outputs
📊 Dashboard
⚡ API
🔔 Alerts
📑 Reports
0 Models Ranked by Arena ELO
0 Models with Live Token Pricing
0 Benchmarks Compared Head-to-Head
0 Original Indices Published
0 Refresh Cadence

Where the numbers come from

Pro hasn't launched, so there are no customer stories to tell yet. Here's the thing we'd rather show you instead: every source, named, so you can check our work.

Arena rankings and head-to-head ELO come from LMArena, the community leaderboard that ranks models by blind pairwise human preference.
LM
LMArena Rankings & ELO
Per-model token pricing comes from OpenRouter, which publishes live input and output rates across every major provider it routes to.
OR
OpenRouter Pricing
Model metadata and ecosystem activity come from HuggingFace and the GitHub API. The two indices are ours, computed from those inputs and documented on AI Pulse.
HF
HuggingFace & GitHub Metadata & activity
Built by victorhugo.ai

Simple pricing

Explorer

$ 0 /mo

Everything on dogkitchen.io, free forever

  • AI Pulse — current rankings & costs
  • Cost Efficiency & Parity indices
  • AI news feed with quality scoring
  • Resource index (300+)
  • Influencer rankings, Learn & Prompt Lab
Get Started

Enterprise

Custom

For organizations with complex needs

  • Everything in Pro
  • Unlimited API calls
  • Custom data integrations
  • Parquet export & warehouse sync
  • SSO & SAML
  • Dedicated account manager
  • SLA guarantees
  • On-prem deployment option
  • Custom onboarding
Contact Sales

Get in early

Pro is still being built. Tell us which part you'd actually use and we'll build that first — and you'll be first in when it ships.

🔒 We respect your data. No spam, ever.