Methodology

Download data

We query the public Hugging Face Hub API nightly and store a snapshot of every catalog model's trailing-30-day downloads, all-time downloads and likes at 04:00 UTC. Hugging Face shows only the current trailing-30-day figure; the day-by-day history — trend charts, 7-day growth rankings, crossover dates — exists because we record it ourselves, daily, since 28 Jul 2026. The catalog covers roughly the top 1,000 models, admitted per task category by fixed quotas so a text-generation giant never crowds out the top speech or embedding models. Known CI/test repositories with inflated download counters are excluded.

VRAM estimates

est_vram_gb = file_size_gb × 1.1
            + kv_cache(context, params)   # 0.15 GB per B params at 8K ctx, linear in ctx
            + 0.5 GB runtime overhead

For GGUF files the quantization is already reflected in the file size, so the file size is the input. For unquantized checkpoints we use parameter count × 2 bytes (fp16). The KV-cache term is a deliberate simplification — a linear approximation calibrated at 8K context — because the honest alternative (per-architecture head math) changes the answer by less than the runtime variance between llama.cpp and vLLM. Verdicts against a GPU: comfortable below 82% of usable memory, tight up to 100%, won't fit above. Apple-silicon presets count 75% of unified memory as usable. These are estimates, not benchmarks; measured tokens/sec is a future, crowdsourced feature.

Open weights vs API-only, pricing and benchmarks

Hosted-API data comes from the public OpenRouter API, imported daily at 05:00 UTC. Every model carries a deployment type: open weights when OpenRouter's listing points to a published Hugging Face repository we can verify (the model is self-hostable, and may also be served via API), and API-only when there is no public repository — the model can only be called through a hosted API. Aliases ("latest" pointers), routers and the :free/:batch rows of the same model are folded into one entry; ambiguous cases are reviewed by hand rather than matched by name.

price_per_1M   = OpenRouter price per token × 1,000,000   # USD, list price
blended_price  = (3 × input + 1 × output) / 4             # typical 3:1 prompt:completion mix
cost_1M_pair   = input + output                            # "1M in + 1M out" rows
value          = AA intelligence index / blended_price     # "Best value via API"

Prices are list prices of the default route; time-of-day and long-context tiers are shown separately in plain words, and per-provider prices come from OpenRouter's endpoint listing. A price snapshot is stored once per model per day, so price-history charts appear after 7 days of data. Intelligence, coding and agentic indices are Artificial Analysis figures and Design Arena ratings as published through OpenRouter — we display them, we do not measure them. API comparisons pair models with similar intelligence index (within 8 points, and within 5× in blended price; within 5 points for open-vs-closed pages). Attribution: pricing data © OpenRouter, benchmark indices © Artificial Analysis.

Licences

The "commercial use" badge is a three-way classification of the repository's declared licence: permissive licences (Apache-2.0, MIT, BSD, CC-BY…) map to allowed; non-commercial licences (CC-BY-NC and variants) to not allowed; community licences with usage conditions (Llama, Gemma, OpenRAIL, FLUX…) to conditional. It is a signpost, not legal advice — always read the licence text linked on the model page.

Verdicts and prompt outputs

Comparison verdicts are generated by a language model from structured data only — the two models' specs, benchmark results, licence classes, VRAM estimates and download trends. The generator is not allowed to use its own knowledge about the models, which prevents confident hallucination. Prompt-gallery outputs are real completions produced once through Hugging Face inference providers, cached permanently, and never edited or cherry-picked. Both carry an "auto-generated" label wherever they appear.

Data sources & terms

• Hugging Face Hub API — model metadata, licences, file inventories, download and like counters for open-weights models. Our daily snapshots of those counters are our own dataset.
• OpenRouter API — API-only models, prices per 1M tokens, :free/:batch variants, provider endpoints, release dates, knowledge cutoffs and retirement dates.
• Artificial Analysis (intelligence, coding and agentic indices) and Design Arena ratings — third-party benchmarks surfaced via OpenRouter. We display them; we do not run them.

Derived figures (VRAM estimates, trends, value scores, licence classes) are ours and documented on this page. When you reuse data from the site, the JSON API or the MCP server, credit “Data: Hugging Face API, OpenRouter API, Artificial Analysis and Design Arena via OpenRouter. Compiled by aimodelscomparison.com” and link the model's page. Provider data remains subject to its source's terms; every value is a best-effort copy refreshed daily, not a guarantee of current list prices.