# Metered: LLM API pricing and AI gateways

> Input, output, cached and batch prices per million tokens for 253 language models across 66 providers, with dated price history, and pricing for 22 AI gateways. By fru.dev.

Each model has its lab's own list price and the price on every other route that sells it (AWS Bedrock, Azure, Google Vertex AI, and hosts like Together, Fireworks, Groq, DeepInfra), read from OpenRouter's public model catalog and endpoint list every day at 05:00 UTC. Price history before tracking began comes from Internet Archive snapshots of that catalog; each change is dated by the first snapshot showing the new price. Gateway pricing (markup, bring-your-own-key fees, free tiers, plans) was read by hand from each gateway's own pricing page, and those pages are re-checked every Monday. Quality scores come from leadersboard.fru.dev. Last successful refresh: 2026-09-24 05:09:46 UTC. 151 dated list-price changes on file.

Disclaimer: List prices from providers' pages on the date shown; your contract may differ. Logos via logo.dev; trademarks belong to their owners. Corrections: use Suggest a correction on any model or gateway page.

## Pages

- [Home: the price map](https://metered.fru.dev/)
- [All models](https://metered.fru.dev/models)
- [Cost calculator](https://metered.fru.dev/calculator): monthly bill by model, with caching, batch and gateway markups
- [AI gateways compared](https://metered.fru.dev/gateways)
- [Price changes timeline](https://metered.fru.dev/changes)
- [Providers](https://metered.fru.dev/providers): [OpenAI](https://metered.fru.dev/providers/openai), [Anthropic](https://metered.fru.dev/providers/anthropic), [Google Gemini API](https://metered.fru.dev/providers/google), [Mistral AI](https://metered.fru.dev/providers/mistral-ai), [Cohere](https://metered.fru.dev/providers/cohere), [xAI](https://metered.fru.dev/providers/xai), [DeepSeek](https://metered.fru.dev/providers/deepseek), [Meta](https://metered.fru.dev/providers/meta), [Qwen (Alibaba)](https://metered.fru.dev/providers/qwen), [Moonshot AI](https://metered.fru.dev/providers/moonshot-ai), [Z.ai](https://metered.fru.dev/providers/z-ai), [Amazon Nova](https://metered.fru.dev/providers/amazon)
- [Sources and method](https://metered.fru.dev/about)
- [Compare gpt-6-sol vs claude-opus-5](https://metered.fru.dev/compare/gpt-6-sol-vs-claude-opus-5)
- [Compare gpt-6-sol vs claude-sonnet-5](https://metered.fru.dev/compare/gpt-6-sol-vs-claude-sonnet-5)
- [Compare gpt-6-astra vs claude-fable-5-1](https://metered.fru.dev/compare/gpt-6-astra-vs-claude-fable-5-1)
- [Compare gpt-6-luna vs gemini-3-8-flash](https://metered.fru.dev/compare/gpt-6-luna-vs-gemini-3-8-flash)
- [Compare claude-sonnet-5 vs gemini-3-8-flash](https://metered.fru.dev/compare/claude-sonnet-5-vs-gemini-3-8-flash)
- [Compare claude-opus-5-5 vs gpt-6-sol-pro](https://metered.fru.dev/compare/claude-opus-5-5-vs-gpt-6-sol-pro)

## Top-rated models (leadersboard.fru.dev score), per 1M tokens

- [Claude Fable 5.1](https://metered.fru.dev/models/claude-fable-5-1): $10 in, $50 out, 1M context
- [GPT-6 Astra](https://metered.fru.dev/models/gpt-6-astra): $10 in, $50 out, 1.05M context
- [Claude Opus 5](https://metered.fru.dev/models/claude-opus-5): $5 in, $25 out, 1M context
- [Claude Fable 5](https://metered.fru.dev/models/claude-fable-5): $10 in, $50 out, 1M context
- [GPT-5.6 Sol](https://metered.fru.dev/models/gpt-5-6-sol): $2 in, $10 out, 1.05M context
- [Claude Opus 4.6](https://metered.fru.dev/models/claude-opus-4-6): $5 in, $25 out, 1M context
- [Claude Opus 4.7](https://metered.fru.dev/models/claude-opus-4-7): $5 in, $25 out, 1M context
- [Muse Spark 1.3](https://metered.fru.dev/models/muse-spark-1-3): $1.25 in, $4.25 out, 1M context
- [Claude Opus 5.5](https://metered.fru.dev/models/claude-opus-5-5): $4 in, $20 out, 1M context
- [GPT-5.5](https://metered.fru.dev/models/gpt-5-5): $5 in, $30 out, 1.05M context
- [GPT-5.4 Pro](https://metered.fru.dev/models/gpt-5-4-pro): $30 in, $180 out, 1.05M context
- [Gemini 3.8 Flash](https://metered.fru.dev/models/gemini-3-8-flash): $0.75 in, $3.75 out, 1M context

## Cheapest models (blended 3:1)

- [Qwen3 Embedding 8B](https://metered.fru.dev/models/qwen3-embedding-8b): $0.01 in, Free out
- [Text Embedding 3 Small](https://metered.fru.dev/models/text-embedding-3-small): $0.02 in, Free out
- [Qwen3 Embedding 4B](https://metered.fru.dev/models/qwen3-embedding-4b): $0.02 in, Free out
- [Mistral Nemo](https://metered.fru.dev/models/mistral-nemo): $0.018 in, $0.03 out
- [Llama 3.1 8B Instruct](https://metered.fru.dev/models/llama-3-1-8b-instruct): $0.02 in, $0.04 out
- [gpt-oss-20b](https://metered.fru.dev/models/gpt-oss-20b): $0.018 in, $0.09 out
- [Qwen3.7 Flash](https://metered.fru.dev/models/qwen3-7-flash): $0.03 in, $0.13 out
- [Mistral Small 3](https://metered.fru.dev/models/mistral-small-24b-instruct-2501): $0.05 in, $0.08 out

## Ways to buy model access (gateways, cloud and data platforms, developer hubs)

- [Helicone](https://metered.fru.dev/gateways/helicone) (independent): 0% markup on credits; plans by requests and seats (Pro $79 a month), BYOK, self-hostable
- [Kong AI Gateway](https://metered.fru.dev/gateways/kong-ai-gateway) (independent): $100 a month per model proxied (Konnect Plus); no token fee, self-hostable
- [LiteLLM](https://metered.fru.dev/gateways/litellm) (independent): Open source, self-hosted with your own keys; enterprise license by quote, BYOK, self-hostable
- [Vercel AI Gateway](https://metered.fru.dev/gateways/vercel-ai-gateway) (independent): Provider list prices, no markup or platform fee on tokens, BYOK
- [Cloudflare AI Gateway](https://metered.fru.dev/gateways/cloudflare-ai-gateway) (independent): Core features free; 5% fee on Unified Billing credits, BYOK
- [Requesty](https://metered.fru.dev/gateways/requesty) (independent): 5% on top of model costs, pay as you go, BYOK
- [Eden AI](https://metered.fru.dev/gateways/eden-ai) (independent): Provider prices passed through; 5.5% fee on credit purchases, BYOK
- [OpenRouter](https://metered.fru.dev/gateways/openrouter) (independent): Provider prices passed through; 5.5% fee on card credit purchases (5% crypto), BYOK
- [Martian Gateway](https://metered.fru.dev/gateways/martian) (independent): No public pricing
- [Portkey](https://metered.fru.dev/gateways/portkey) (independent): Per logged request: free 10k logs a month, Production from $49 a month, BYOK, self-hostable
- [TrueFoundry AI Gateway](https://metered.fru.dev/gateways/truefoundry) (independent): Per user and per request ($25 per user a month on Pro), self-hostable
- [AWS Bedrock](https://metered.fru.dev/gateways/aws-bedrock) (cloud): Pass-through per 1M tokens at the lab's list price for Global cross-region inference; Geo and in-region inference cost 10% more; billed on your AWS account
- [Azure AI Foundry](https://metered.fru.dev/gateways/azure-ai-foundry) (cloud): Per 1M (or 1K) tokens on your Azure bill; Global deployment at OpenAI list price, Data Zone and Regional deployments cost 10% more; PTUs for provisioned throughput
- [Google Vertex AI](https://metered.fru.dev/gateways/google-vertex-ai) (cloud): Per 1M tokens on your Google Cloud bill; partner models at the lab's list price on the Global endpoint, regional/multi-region endpoints about 10% more
- [IBM watsonx.ai](https://metered.fru.dev/gateways/ibm-watsonx-ai) (cloud): Per 1M tokens (billed as Resource Units, 1 RU = 1,000 tokens incl. input and output); Essentials plan from USD 0/month, Standard from USD 1110/month
- [OCI Generative AI](https://metered.fru.dev/gateways/oci-generative-ai) (cloud): Mixed: per 1M tokens for Google, xAI and OpenAI gpt-oss models; per 10,000 transactions (1 transaction = 1 character of prompt plus response) for Meta and Cohere; dedicated AI clusters per unit hour
- [Databricks Mosaic AI](https://metered.fru.dev/gateways/databricks-mosaic-ai) (data): DBUs per 1M tokens (Standard pay per token; Priority and Batch tiers separate)
- [Snowflake Cortex AI](https://metered.fru.dev/gateways/snowflake-cortex-ai) (data): Snowflake AI Credits per 1M tokens (input and output priced separately)
- [BigQuery ML](https://metered.fru.dev/gateways/bigquery-ml-genai) (data): BigQuery bytes processed (on-demand or editions) plus the underlying Vertex AI model charges billed directly by that service; no separate BigQuery per-token rate for LLM calls
- [Hugging Face Inference](https://metered.fru.dev/gateways/huggingface-inference-providers) (hub): Pass-through at each provider's rate with no Hugging Face markup; pay as you go on your HF account, or bill directly to the provider with your own key, BYOK
- [Cloudflare Workers AI](https://metered.fru.dev/gateways/cloudflare-workers-ai) (hub): Neurons: $0.011 per 1,000 Neurons, with per-model token prices published as the Neuron equivalent
- [NVIDIA NIM](https://metered.fru.dev/gateways/nvidia-nim) (hub): Hosted API endpoints free for NVIDIA Developer Program prototyping; production self-hosted NIM requires an NVIDIA AI Enterprise license from $4,500 per GPU per year (about $1 per GPU per hour in the cloud); no per-token price

## Recent list-price changes

- 2026-09-24: DeepSeek V4 Pro 0813 cache_read $0.022 to $0.044
- 2026-09-24: DeepSeek V4 Pro 0813 output $1.98 to $3.96
- 2026-09-24: DeepSeek V4 Pro 0813 input $0.66 to $1.32
- 2026-09-24: DeepSeek V4.1 Flash cache_read $0.003 to $0.006
- 2026-09-24: DeepSeek V4.1 Flash output $0.6 to $1.2
- 2026-09-24: DeepSeek V4.1 Flash input $0.15 to $0.3
- 2026-08-29: Gemini 3.7 Flash cache_read $0.037 to $0.075
- 2026-08-29: Gemini 3.7 Flash output $1.88 to $3.75

## Data

- [Every current price as plain text](https://metered.fru.dev/llms-full.txt)
- [Sitemap](https://metered.fru.dev/sitemap.xml)
- [JSON: all models with headline prices](https://metered.fru.dev/api/models)
- JSON: one model with every route and its history: https://metered.fru.dev/api/models/<id>
- [JSON: gateways](https://metered.fru.dev/api/gateways)
- [JSON: list-price changes](https://metered.fru.dev/api/changes)