# LLM Pricing

> Pricing, quality leaderboards and selection tools for 2537+ LLM/chat models across 193 providers. Data from models.dev and Artificial Analysis.

## Overview

- Models: 2537
- Providers: 193
- Provider quotes: 7219
- Data synced: 2026-08-21

## About

LLM Pricing aggregates and normalizes API pricing, capabilities and quality benchmarks for large language / chat models, so you can compare price, capability and every provider's quote in one place.

Methodology:
- Models are merged into canonical entries by normalized name (a model hosted on many providers is one entry with per-provider quotes).
- Blended unit price = input x 0.75 + output x 0.25 (approximate typical usage mix).
- Providers are tiered: first-party (the lab itself), cloud platforms, and third-party gateways. The lowest price is taken from paid channels; $0 channels are listed separately (free tiers usually have rate limits).
- Capability flags are tri-state: supported / unsupported / unknown — never conflated.

## Data sources

- Pricing & capabilities: models.dev (MIT licensed).
- Quality, coding, agentic, speed & cost-per-task scores: Artificial Analysis.
- Real-world usage (30-day tokens): OpenRouter.
- Data is refreshed regularly (last sync 2026-08-21); prices are indicative — always confirm on each provider's official page.

## Key pages

- [Pricing](https://llmpricing.dev/en/): compare price, capability and every provider's quote for all models
- [Leaderboards](https://llmpricing.dev/en/leaderboards/): intelligence, coding, value, cost per task, speed, usage
- [Value frontier](https://llmpricing.dev/en/frontier/): best intelligence per dollar
- [Providers](https://llmpricing.dev/en/providers/): first-party, cloud and gateway providers
- [Compare](https://llmpricing.dev/en/compare/): 2-4 models side by side
- [Releases](https://llmpricing.dev/en/releases/): a dated feed of recent model releases, each linking to the model's page

## Tools

- [Cost calculator](https://llmpricing.dev/en/tools/cost/): estimate monthly API cost across providers from token usage and cache hit rate
- [Token counter](https://llmpricing.dev/en/tools/token/): exact token counts for OpenAI tokenizers (o200k/cl100k) plus labeled estimates for other models, and per-model cost
- [Subscription vs API](https://llmpricing.dev/en/tools/subscription/): whether an AI subscription beats pay-as-you-go for your usage (ranges, not single points)
- [AI subscription prices](https://llmpricing.dev/en/subscriptions/): consumer plans and quotas for ChatGPT, Claude, Gemini, Grok, Perplexity, Copilot and more

## Data & API

Normalized data is published as free static JSON (CC BY 4.0, no key required):
- [Models JSON](https://llmpricing.dev/api/models.json): every model with prices, blended unit price, context, capabilities, quality scores and usage
- [Providers JSON](https://llmpricing.dev/api/providers.json): every provider with tier, model count, SDK and API base
- [API & MCP docs](https://llmpricing.dev/en/api/): field reference, code examples and how to consume the data from an MCP-enabled agent

## Popular models

- [GLM-5.2](https://llmpricing.dev/m/zhipuai/glm-5.2/) — Z.AI, $1.40/$4.40 per 1M, 84 providers
- [GPT OSS 120B](https://llmpricing.dev/m/openai/gpt-oss-120b/) — Regolo AI, $1.00/$4.20 per 1M, 69 providers
- [Kimi K2.6](https://llmpricing.dev/m/moonshotai/kimi-k2.6/) — Moonshot AI, $0.95/$4.00 per 1M, 65 providers
- [DeepSeek V4 Pro](https://llmpricing.dev/m/deepseek/deepseek-v4-pro/) — DeepSeek, $0.435/$0.87 per 1M, 60 providers
- [DeepSeek V4 Flash](https://llmpricing.dev/m/deepseek/deepseek-v4-flash/) — DeepSeek, $0.14/$0.28 per 1M, 58 providers
- [Kimi K2.7 Code](https://llmpricing.dev/m/moonshotai/kimi-k2.7-code/) — Moonshot AI, $0.95/$4.00 per 1M, 57 providers
- [Kimi K3](https://llmpricing.dev/m/moonshotai/kimi-k3/) — Moonshot AI, $3.00/$15.00 per 1M, 56 providers
- [GPT-5.5](https://llmpricing.dev/m/openai/gpt-5.5/) — OpenAI, $5.00/$30.00 per 1M, 45 providers
- [Claude Sonnet 4.6](https://llmpricing.dev/m/anthropic/claude-sonnet-4-6/) — Anthropic, $3.00/$15.00 per 1M, 44 providers
- [GPT-5.4](https://llmpricing.dev/m/openai/gpt-5.4/) — OpenAI, $2.50/$15.00 per 1M, 42 providers
- [MiniMax-M2.7](https://llmpricing.dev/m/minimax/MiniMax-M2.7/) — MiniMax (minimaxi.com), $0.30/$1.20 per 1M, 42 providers
- [Claude Opus 4.7](https://llmpricing.dev/m/anthropic/claude-opus-4-7/) — Anthropic, $5.00/$25.00 per 1M, 41 providers
- [Claude Opus 4.8](https://llmpricing.dev/m/anthropic/claude-opus-4-8/) — Anthropic, $5.00/$25.00 per 1M, 41 providers
- [DeepSeek V4 Flash 0731](https://llmpricing.dev/m/deepseek/deepseek-v4-flash-0731/) — Charm Hyper, $0.479/$1.44 per 1M, 40 providers
- [Claude Opus 4.6](https://llmpricing.dev/m/anthropic/claude-opus-4-6/) — Anthropic, $5.00/$25.00 per 1M, 38 providers
- [MiniMax-M3](https://llmpricing.dev/m/minimax/MiniMax-M3/) — MiniMax (minimaxi.com), $0.30/$1.20 per 1M, 38 providers
- [GPT-5.4 mini](https://llmpricing.dev/m/openai/gpt-5.4-mini/) — OpenAI, $0.75/$4.50 per 1M, 36 providers
- [GPT-5.6 Sol](https://llmpricing.dev/m/openai/gpt-5.6-sol/) — OpenAI, $5.00/$30.00 per 1M, 34 providers
- [GPT-5.6 Luna](https://llmpricing.dev/m/openai/gpt-5.6-luna/) — OpenAI, $0.20/$1.20 per 1M, 34 providers
- [GPT-5.6 Terra](https://llmpricing.dev/m/openai/gpt-5.6-terra/) — OpenAI, $2.00/$12.00 per 1M, 34 providers
- [GPT-5.2](https://llmpricing.dev/m/openai/gpt-5.2-2/) — OpenAI, $1.75/$14.00 per 1M, 33 providers
- [Claude Sonnet 5](https://llmpricing.dev/m/anthropic/claude-sonnet-5/) — Anthropic, $2.00/$10.00 per 1M, 32 providers
- [Gemini 2.5 Flash](https://llmpricing.dev/m/google/gemini-2.5-flash/) — Vertex, $0.30/$2.50 per 1M, 31 providers
- [Gemini 3.1 Pro Preview](https://llmpricing.dev/m/google/gemini-3.1-pro-preview/) — Vertex, $2.00/$12.00 per 1M, 31 providers
- [GPT-5 Mini](https://llmpricing.dev/m/openai/gpt-5-mini/) — OpenAI, $0.25/$2.00 per 1M, 31 providers
- [Claude Fable 5](https://llmpricing.dev/m/anthropic/claude-fable-5/) — Anthropic, $10.00/$50.00 per 1M, 31 providers
- [GPT-5.4 nano](https://llmpricing.dev/m/openai/gpt-5.4-nano/) — OpenAI, $0.20/$1.25 per 1M, 30 providers
- [Qwen3.7 Max](https://llmpricing.dev/m/alibaba/qwen3.7-max/) — Alibaba (China), $2.50/$7.50 per 1M, 30 providers
- [Gemini 2.5 Pro](https://llmpricing.dev/m/google/gemini-2.5-pro/) — Vertex, $1.25/$10.00 per 1M, 29 providers
- [Gemini 3.5 Flash](https://llmpricing.dev/m/google/gemini-3.5-flash/) — Vertex, $1.50/$9.00 per 1M, 29 providers
- [GPT-5](https://llmpricing.dev/m/openai/gpt-5-2/) — OpenAI, $1.25/$10.00 per 1M, 28 providers
- [GPT-5.1](https://llmpricing.dev/m/openai/gpt-5.1/) — OpenAI, $1.25/$10.00 per 1M, 28 providers
- [Claude Opus 5](https://llmpricing.dev/m/anthropic/claude-opus-5/) — Anthropic, $5.00/$25.00 per 1M, 28 providers
- [Qwen3.7 Plus](https://llmpricing.dev/m/alibaba/qwen3.7-plus/) — Alibaba (China), $0.50/$3.00 per 1M, 28 providers
- [GPT-5.3 Codex](https://llmpricing.dev/m/openai/gpt-5.3-codex/) — OpenAI, $1.75/$14.00 per 1M, 27 providers
- [Qwen3.6 Plus](https://llmpricing.dev/m/alibaba/qwen3.6-plus/) — Alibaba (China), $0.50/$3.00 per 1M, 27 providers
- [DeepSeek V4 Pro 0813](https://llmpricing.dev/m/deepseek/deepseek-v4-pro-0813/) — Venice AI, $1.65/$4.95 per 1M, 26 providers
- [GPT-5 Nano](https://llmpricing.dev/m/openai/gpt-5-nano/) — OpenAI, $0.05/$0.40 per 1M, 26 providers
- [Qwen3.6 27B](https://llmpricing.dev/m/alibaba/qwen3.6-27b/) — Alibaba, $0.60/$3.60 per 1M, 26 providers
- [Qwen3.6 35B-A3B](https://llmpricing.dev/m/alibaba/qwen3.6-35b-a3b/) — Alibaba, $0.248/$1.49 per 1M, 25 providers