The AI tooling landscape changes weekly. A model that was cheapest in January costs twice as much by March. Capabilities shift. New providers launch and disappear. Pricing pages bury the real numbers under enterprise-speak and asterisks.
For developers, founders, and teams making real AI spend decisions, this creates a genuine problem: where do you find current, neutral, comparable information?
Most existing sources fall short. Vendor sites are optimized to sell, not inform. General comparison sites lack the depth for AI-specific decision-making. Reddit threads are valuable but unstructured, quickly stale, and impossible to compare across providers.
I built AI Cost & Tools Hub to fill this gap. The site tracks pricing for 264 LLM models across 20 providers, with every number verified manually against the vendor's own canonical pricing page. The methodology for every number is published. When pricing changes, I re-verify and stamp a new date. When a data point isn't yet verified, the field stays blank rather than fabricated.
What's live today: per-model pricing pages with input / output / cached rates, blended-cost scenarios, price history, and a provider-hub for each frontier vendor. What's next: per-use-case cost calculators and side-by-side comparisons (Phase 2 and 3 of the roadmap).
I accept no vendor sponsorships for editorial content. AI companies cannot pay to appear in comparisons, improve rankings, or influence review scores. Revenue comes from disclosed affiliate commissions and display advertising — not from the companies covered.
Every number published has a traceable source. Pricing data links to the provider's official pricing page. Benchmark scores cite the test conditions, date, and model version used. The full methodology is at /methodology/.
When a model is expensive for its capability tier, I say so. When a pricing page is deceptive, I analyze the deception. When a tool is not worth the cost, I recommend against it even when affiliate commission is on the table. Losing affiliate revenue for telling the truth is a feature, not a bug.
When a number is wrong, I want to hear it. Reader corrections at [email protected] — pricing changes, factual errors, methodology bugs — get published at /corrections/ with the original claim, the fix, and the date. No silent edits. The site improves only as fast as readers catch what I miss.
Sponsored content is labeled "Sponsored" in the headline, the URL slug, and the schema markup. It never appears in organic comparison tables or recommendation widgets. I do not let sponsored arrangements affect editorial rankings — the wall of separation is enforced as individual discipline, not by a separate team.
I use AI tools for drafting, formatting, and data structuring. Every published claim, recommendation, and comparison is reviewed and signed off by a named human author — me. AI-generated text that has not been reviewed by a human does not ship.
Founder of AI Cost & Tools Hub. Editorial sign-off on every pricing claim — verified against the vendor's own canonical pricing page on the date stamped at the bottom of that page.
See profileI'm open to collaboration with vendors, agencies, researchers, and platforms working in the AI tooling space.
If you have an idea for a sponsorship, data partnership, integration, co-published research, or anything else that could be useful to this audience — send a proposal and I'll get back to you within 5 business days.
Long-form pieces clearly labeled "Sponsored" — produced to your brief. Never inserted into rankings or comparison tables.
License the pricing and benchmark feeds. Co-build datasets. Embed the calculators inside your product or docs.
Joint benchmarks, market reports, or surveys with your team — I handle methodology, you bring expertise or data.
Banner inventory on high-intent comparison and pricing pages. Always labeled, never mixed into editorial rankings.
Tell me who you are, what you're proposing, audience size if relevant, and a rough timeline. No deck required — a short, honest email works best.
When a pricing figure, benchmark score, or factual claim is incorrect, the page gets updated, the change is noted inline, and an entry is added to the public corrections log. No silent edits. The original claim and the corrected claim are both recorded.
Corrections logUse the "Report an error" link at the bottom of any page, or email me directly at [email protected]. Include the page URL, the incorrect claim, and a source for the correct information. I aim to review every report within 48 hours.
Report an errorEvery correction issued is recorded on the corrections page with the original claim, the corrected claim, and the date. Rather than quote a static count here that could drift out of date, the live log is the canonical source of truth.
View corrections logThis site operates in build-in-public mode. The methodology changelog is public and linkable. Active open questions about data sourcing, scoring models, and editorial standards are documented and linked below. If you have an opinion on how a tricky methodological question should be handled, I want to hear it — email [email protected].
No. Rankings and recommendations in editorial content are determined solely by the data. Vendors cannot pay to influence their position. Clearly labeled sponsored content may appear on the site under a separate sponsorship workflow — but it never appears in comparison tables or recommendation widgets, and the wall of separation is enforced as individual discipline by the founder, not by a separate team. See the sponsorship policy.
I verify each price manually against the vendor's own canonical pricing page using WebFetch — no scrapers, no third-party aggregators, no billing-API integrations. Updates land within 24 hours of any vendor change I detect. Every data point links to the source it was verified against. See the full methodology.
All editorial content is written and signed off by the founder. AI tools assist with drafting and formatting; no AI-generated content is published without human review and sign-off.
Pricing parsers run daily. When a price changes, the page is updated automatically, a changelog entry is created, and a human reviewer verifies the change before it goes live on the page. Historical pricing is preserved in the changelog.
Yes. Data licensing is available for teams who need structured AI pricing data via API or data exports. Contact [email protected] for details and pricing.
Use the "Report an error" link at the bottom of any page, or email [email protected] with the page URL, the incorrect claim, and a source for the correct information. I aim to review within 48 hours and log all corrections publicly.