Open data
Updated
BroadRouter publishes seven open datasets: tokenizer efficiency, real cost by language, an LLM API price index, the raw provider price database behind it, LLM API price history from archived pricing pages, speech-to-text API prices, and LLM API free tiers and rate limits. Each has a method section and CSV and JSON downloads, under CC BY 4.0.
Tokenizer efficiency across languages: 16 tokenizers from DeepSeek, GLM, Qwen, Gemma, OpenAI, Meta, Moonshot, MiniMax, Microsoft, Mistral and NVIDIA
Token counts for the same 2,009 FLORES-200 sentences in Modern Standard Arabic, four Arabic dialects, Hindi, Urdu, Chinese, French, Spanish and English, plus 41 openly licensed texts (Gulf Arabic and Arabizi sentences, Wikipedia prose, code and JSON), with 95% intervals, under 16 official tokenizers (DeepSeek, GLM, Qwen, Gemma, GPT, gpt-oss, Llama, Kimi, MiniMax, Phi, Mistral, Nemotron). English prose: about 1,290 tokens per 1,000 words.
- Licence
- CC BY 4.0 for our figures. The source texts in the JSON download are not relicensed: each keeps its source licence, given per text.
- Method
- Measured with pinned tokenizer files by a committed script; parallel sentences from FLORES-200 (CC BY-SA 4.0, counts only, tarball pinned by SHA-256) and Tatoeba, other texts from Wikipedia, Project Gutenberg and Wikidata, each with its source and licence.
- Updates
- Re-measured when a tokenizer or the corpus changes; costs recomputed on every build from BroadNet pricing.json.
- Measured
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). Tokenizer efficiency across languages: 16 tokenizers from DeepSeek, GLM, Qwen, Gemma, OpenAI, Meta, Moonshot, MiniMax, Microsoft, Mistral and NVIDIA (version 4.0) [Data set]. BroadNet Technologies. https://broadrouter.com/data/tokenizer-efficiency/@misc{broadrouter_tokenizer_efficiency, author = {BroadRouter}, title = {Tokenizer efficiency across languages: 16 tokenizers from DeepSeek, GLM, Qwen, Gemma, OpenAI, Meta, Moonshot, MiniMax, Microsoft, Mistral and NVIDIA}, year = {2026}, version = {4.0}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/tokenizer-efficiency/}}, note = {Data set, data as of 2026-11-15}, }Real cost by language: what the same text costs on each LLM model and provider
What the same text costs in English, Arabic and its dialects, Hindi, Urdu, Chinese, French, Spanish, code and JSON: input and output cost per 1M characters and per 1,000 words for 618 standard-tier prices (108 models, 35 providers plus BroadNet), from measured tokenizers and providers' own prices.
- Licence
- CC BY 4.0 for our figures. Each price is a fact published by the provider named in its row, at the source URL given per row; we do not license those pages to you.
- Method
- Token counts from the tokenizer efficiency dataset; each model matched to its tokenizer by a script that checks the model's own repo; prices from providers' own pages and BroadNet's pricing.json.
- Updates
- Costs recomputed on every build from the latest prices; model matching re-checked when the script is re-run.
- Checked
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). Real cost by language: what the same text costs on each LLM model and provider (version 1.1) [Data set]. BroadNet Technologies. https://broadrouter.com/data/real-cost-by-language/@misc{broadrouter_real_cost_by_language, author = {BroadRouter}, title = {Real cost by language: what the same text costs on each LLM model and provider}, year = {2026}, version = {1.1}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/real-cost-by-language/}}, note = {Data set, data as of 2026-11-15}, }LLM API price index: prices from providers' own pages, by model, provider and context length
1671 prices for 428 models from 41 providers, each read from the provider's own pricing page. Median, cheapest and dearest price per model, provider spread, maker versus host, batch and flex discounts, output-to-input ratios, cached pricing and context tiers. Median blended price: USD 0.6581 per 1M tokens.
- Licence
- CC BY 4.0 for our figures. Each price is a fact published by the provider named in its row, at the source URL given per row; we do not license those pages to you.
- Method
- Read from each provider's own pricing page, with source URL, check date and a quote per price; statistics computed at build time. No aggregator data; listed prices, not BroadNet traffic.
- Updates
- Source pages re-fetched weekly and flagged changes re-read with AI assistance; statistics recomputed on every build.
- Checked
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). LLM API price index: prices from providers' own pages, by model, provider and context length (version 2026-11-15) [Data set]. BroadNet Technologies. https://broadrouter.com/data/llm-api-price-index/@misc{broadrouter_llm_price_index, author = {BroadRouter}, title = {LLM API price index: prices from providers' own pages, by model, provider and context length}, year = {2026}, version = {2026-11-15}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/llm-api-price-index/}}, note = {Data set, data as of 2026-11-15}, }LLM API provider price database: every price with its source, date and quote
The raw database behind the price index: 1,671 LLM API prices for 430 models from 41 providers, one row per price, with provider model ID, service tier, input, cached and output price, context, precision, source URL, date checked and a verbatim quote.
- Licence
- CC BY 4.0 for our figures. Each price is a fact published by the provider named in its row, at the source URL given per row, and each quote is a short excerpt from that page; we do not license those pages to you.
- Method
- Read from each provider's own pricing page with AI assistance and a page renderer; every row keeps its source URL, check date and quote. No aggregator data; BroadNet's own prices not included.
- Updates
- Source pages re-fetched weekly; flagged rows re-read and updated, each with its own check date.
- Checked
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). LLM API provider price database: every price with its source, date and quote (version 2026-11-15) [Data set]. BroadNet Technologies. https://broadrouter.com/data/llm-api-provider-prices/@misc{broadrouter_llm_provider_prices, author = {BroadRouter}, title = {LLM API provider price database: every price with its source, date and quote}, year = {2026}, version = {2026-11-15}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/llm-api-provider-prices/}}, note = {Data set, data as of 2026-11-15}, }LLM API price history from archived provider pricing pages
LLM API list prices over time: 3,337 input, output and cached prices for 580 models from 11 providers, read from 308 Internet Archive snapshots of their own pricing pages (2023-01-03 to 2026-11-02). 100 list-price cuts and 52 rises on the same model, each with both archive links.
- Licence
- CC BY 4.0 for our figures. Each price is a fact published by the provider named in its row, on the archived page linked per row, and each quote is a short excerpt from that capture; we do not license those pages to you.
- Method
- Monthly Wayback Machine captures fetched in raw form; per-provider extractors read only prices that appear in the capture, with the archive URL and a quote per row; skipped captures listed with the reason; rows spot-checked with AI assistance against the raw HTML.
- Updates
- A fixed historical snapshot, rebuilt when new captures are added; changes from now on are in the price change log.
- Built
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). LLM API price history from archived provider pricing pages (version 2026-11-15) [Data set]. BroadNet Technologies. https://broadrouter.com/data/llm-api-price-history/@misc{broadrouter_llm_price_history, author = {BroadRouter}, title = {LLM API price history from archived provider pricing pages}, year = {2026}, version = {2026-11-15}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/llm-api-price-history/}}, note = {Data set, data as of 2026-11-15}, }Speech-to-text API price index: batch and streaming price per audio hour
Batch and streaming list prices per audio hour for 26 speech-to-text products from 14 providers, read from each provider's own pages, with billing increment, minimum charge, Arabic support, diarization and free tier. Median batch price: USD 0.20 per hour.
- Licence
- CC BY 4.0 for our figures. Each price, language list and policy is a fact published by the provider named in its row, at the source URL given per row; we do not license those pages to you.
- Method
- Read with AI assistance from each provider's own pricing and docs pages, with source URL and check date per row; BroadNet rows from pricing.json. No aggregator data.
- Updates
- Third-party prices re-read with AI assistance (the check date changes when they are); BroadNet rows and statistics recomputed on every build.
- Checked
- 2026-11-15
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). Speech-to-text API price index: batch and streaming price per audio hour (version 2026-11-15) [Data set]. BroadNet Technologies. https://broadrouter.com/data/speech-to-text-api-prices/@misc{broadrouter_stt_price_index, author = {BroadRouter}, title = {Speech-to-text API price index: batch and streaming price per audio hour}, year = {2026}, version = {2026-11-15}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/speech-to-text-api-prices/}}, note = {Data set, data as of 2026-11-15}, }LLM API free tiers and default rate limits for new accounts
Free tiers, free credit, default RPM, TPM, RPD and concurrency limits, card requirements and free-tier training terms for 14 LLM API providers, from their own documentation. 6 let a new account call the API without paying.
- Licence
- CC BY 4.0 for our figures. Each limit, credit and data-use statement is a fact published by the provider named in its row, at the source URL given per row; we do not license those pages to you.
- Method
- Read with AI assistance from each provider's own docs, pricing and privacy pages, with source URL and check date per row; BroadNet from product.ts.
- Updates
- Re-read with AI assistance; BroadNet row recomputed on every build.
- Checked
- 2026-11-14
- Downloads
- CSV · JSON · Croissant · Markdown
Cite this dataset
BroadRouter (2026). LLM API free tiers and default rate limits for new accounts (version 2026-11-14) [Data set]. BroadNet Technologies. https://broadrouter.com/data/llm-api-free-tiers-and-rate-limits/@misc{broadrouter_llm_free_tiers, author = {BroadRouter}, title = {LLM API free tiers and default rate limits for new accounts}, year = {2026}, version = {2026-11-14}, publisher = {BroadNet Technologies}, howpublished = {\url{https://broadrouter.com/data/llm-api-free-tiers-and-rate-limits/}}, note = {Data set, data as of 2026-11-14}, }
BroadRouter is BroadNet's own site. The price index, the provider price database and the price history use third-party listed prices and exclude BroadNet's own prices (BroadNet's pricing pages have no archive captures); tokenizer costs use BroadNet's published rate card, and the real-cost dataset lists BroadNet's rows next to every other provider's, labelled. The speech-to-text and free-tier datasets include BroadNet's rows, ranked by the same rules as every other provider.