BroadRouter

LLM API price index

Updated

The median text model costs USD 0.658 per 1M tokens (3:1 input to output), across 424 models priced on 41 providers' own pages. The same model's dearest standard price is a median 1.3× its cheapest, and output tokens cost a median 4.0× as much as input. Data: CSV and JSON.

Download CSV Download JSON Croissant metadata Compilation CC BY 4.0 · prices credited to each provider · checked 2026-11-15

Every individual price behind this index, with its source URL, check date and quote, is in the LLM API provider price database.

Source: 1,671 prices read from 41 providers' own pricing pages, each with a link and check date (the JSON lists every one). This is not BroadNet traffic data. BroadRouter is BroadNet's own site, and BroadNet's prices are not in the index.

Key findings

  1. The median text model costs USD 0.658 per 1M tokens (3 input tokens per output token) across 424 models with a paid standard price, read from 41 providers' own pricing pages (checked 2026-11-14 to 2026-11-15); the middle half costs USD 0.263 to USD 2.85.
  2. Model price levels run from USD 0.0315 (Tev1 4B Experimental) to USD 67.50 (GPT-5.5 Pro) per 1M tokens, a 2,143-fold range.
  3. For the same model, the dearest standard price is a median 1.3× the cheapest, across 197 models with paid standard prices from two or more providers; 61 of them (31%) have a gap of 2× or more.
  4. The widest gap is Llama 3.1 8B Instruct: USD 0.025 at DeepInfra against USD 0.418 at Cloudflare Workers AI, 17× apart (blended, per 1M tokens).
  5. Where both exist, a third-party host undercuts the model's maker on 29 of 98 models; the maker is cheapest or tied on 69. The median cheapest host costs 1.0× the maker's price.
  6. DeepInfra is the cheapest provider for 49 of the 197 multi-provider models (28 without a tie).
  7. Batch prices are a median 50% below the same provider's standard price, across 156 batch prices from 9 providers.
  8. Output tokens cost a median 4.0× as much as input tokens; the most common ratio is 4.0×, used by 67 of 424 models (15.8%).
  9. 65.3% of standard prices (791 of 1,212) include a cached-input price, at a median 88% off the normal input price. Only 25.9% state the numeric precision served.
  10. By context window, the median price level is lowest for up to 32K tokens (USD 0.245 per 1M, 20 models) and highest for 256K to under 1M tokens (USD 3.00, 36 models).

How much does an LLM API cost per million tokens?

Each model's median blended standard price, grouped by the largest context window a provider states. 424 models with a paid standard price; models with no stated context are left out of this table.

Context windowModelsMedian inputMedian outputMedian blendedCheapestDearest
Up to 32K20USD 0.161USD 0.51USD 0.245USD 0.0425USD 3.78
32K to 128K53USD 0.19USD 0.6USD 0.28USD 0.0607USD 1.59
128K to 256K82USD 0.244USD 1.03USD 0.451USD 0.09USD 3.50
256K to under 1M36USD 2.00USD 6.00USD 3.00USD 0.23USD 20.00
1M and more56USD 0.75USD 2.83USD 1.36USD 0.0875USD 20.00

How much do prices vary between providers for the same model?

197 models with paid standard prices from two or more providers. Blended price per 1M tokens, widest spread first.

ModelProvidersCheapestMedianDearestSpreadCheapest provider
Llama 3.1 8B Instruct6USD 0.025USD 0.22USD 0.41817×DeepInfra
Qwen-MT Plus2USD 0.375USD 2.03USD 3.699.8×Novita AI
Llama 3.3 70B14USD 0.155USD 0.71USD 1.237.9×DeepInfra
Mistral-Nemo-Instruct-24072USD 0.0293USD 0.127USD 0.2257.7×DeepInfra
Hy34USD 0.231USD 0.25USD 1.757.6×SiliconFlow
Llama 3 8B Stheno V3.22USD 0.05USD 0.211USD 0.3727.5×Novita AI
DeepSeek V4 Flash 073115USD 0.09USD 0.33USD 0.667.3×DeepInfra, Hyperbolic
DeepSeek R1 Distill Llama 70B2USD 0.375USD 1.54USD 2.707.2×GMI Cloud
OpenAI GPT OSS 120B21USD 0.065USD 0.263USD 0.456.9×CoreWeave Forge
GPT OSS 20B10USD 0.036USD 0.0775USD 0.2256.3×AkashML
Mistral Medium 32USD 0.8USD 2.78USD 4.766.0×Google Vertex AI
DeepSeek R1 Distill Qwen 7B2USD 0.125USD 0.417USD 0.715.7×GMI Cloud
DeepSeek R1 Distill Qwen 32B3USD 0.6USD 1.59USD 3.245.4×GMI Cloud
Gemma 3 12B4USD 0.075USD 0.107USD 0.3985.3×DeepInfra, Featherless
Gemma 4 31B12USD 0.16USD 0.205USD 0.8135.1×CoreWeave Forge
DeepSeek V3.17USD 0.425USD 0.464USD 2.165.1×DeepInfra
GPT-5.6 Luna6USD 0.45USD 0.45USD 2.255.0×OpenAI, DigitalOcean, Microsoft Azure, Perplexity
NVIDIA Nemotron 3 Super 120B A12B3USD 0.156USD 0.275USD 0.754.8×DeepInfra
Qwen3.8 27B15USD 0.342USD 1.05USD 1.604.7×Chutes
MiMo v2.54USD 0.175USD 0.205USD 0.84.6×Baidu, Parasail
DeepSeek V33USD 0.438USD 0.463USD 2.004.6×SiliconFlow
Llama 3 8B Instruct3USD 0.0931USD 0.375USD 0.4184.5×Featherless
DeepSeek V3.211USD 0.234USD 0.367USD 1.004.3×GMI Cloud
Qwen3 Coder 480B A35B6USD 0.438USD 0.626USD 1.804.1×SiliconFlow
GLM 5.324USD 1.05USD 2.15USD 4.304.1×DeepInfra
Llama 3.1 Instruct (70B)3USD 0.72USD 0.783USD 2.884.0×Amazon Bedrock
DeepSeek V3 03243USD 0.502USD 0.695USD 2.004.0×GMI Cloud
DeepSeek V4.1 Flash 101716USD 0.147USD 0.525USD 0.5253.6×DeepInfra
Qwen 2.5 7B3USD 0.05USD 0.07USD 0.1773.5×SiliconFlow
Mistral-Small-24B-Instruct-25012USD 0.0575USD 0.131USD 0.2043.5×DeepInfra
DeepSeek R1 05285USD 0.695USD 1.00USD 2.363.4×Featherless
Qwen3-8B2USD 0.06USD 0.131USD 0.2023.4×SiliconFlow
DeepSeek R1 Distill Qwen 14B2USD 0.2USD 0.402USD 0.6033.0×GMI Cloud
Qwen3 VL 235B A22B3USD 0.37USD 0.632USD 1.062.9×DeepInfra
Command R2USD 0.263USD 0.506USD 0.752.9×Cohere
Qwen3.6 35B A3B10USD 0.297USD 0.39USD 0.8442.8×CoreWeave Forge
Llama 4 Scout 17B 16E Instruct6USD 0.15USD 0.287USD 0.4152.8×DeepInfra
GLM 5.219USD 0.872USD 2.15USD 2.372.7×DeepInfra
Mistral Small 3.1 24B Instruct4USD 0.15USD 0.211USD 0.4022.7×Google Vertex AI
Gemma 3 27B6USD 0.1USD 0.145USD 0.2682.7×DeepInfra
DeepSeek R14USD 0.92USD 1.64USD 2.362.6×GMI Cloud, SiliconFlow
Kimi K2.613USD 0.743USD 1.45USD 1.882.5×Chutes
Qwen3 235B A22B Instruct 2507 FP85USD 0.153USD 0.3USD 0.3852.5×GMI Cloud
DeepSeek V4 Pro11USD 1.20USD 2.06USD 3.002.5×GMI Cloud
GLM 5.3 Flash20USD 0.119USD 0.237USD 0.2972.5×GMI Cloud
DeepSeek V4.1 Flash16USD 0.21USD 0.525USD 0.5252.5×DeepInfra
Phi-43USD 0.0875USD 0.0875USD 0.2192.5×DeepInfra, Featherless
Mistral Small 3.2 24B Instruct4USD 0.106USD 0.138USD 0.2632.5×DeepInfra
DeepSeek V4 Flash8USD 0.113USD 0.174USD 0.272.4×DeepInfra
MiniMax M310USD 0.412USD 0.525USD 0.9632.3×CoreWeave Forge
Qwen3 Coder Next3USD 0.29USD 0.407USD 0.6752.3×Parasail
Llama 3.2 11B Vision Instruct3USD 0.16USD 0.206USD 0.3712.3×Amazon Bedrock
NVIDIA Nemotron 3.5 Super VL6USD 0.175USD 0.256USD 0.3882.2×Crusoe Cloud
Llama 3.2 3B Instruct4USD 0.122USD 0.136USD 0.2632.2×Featherless
Qwen3 Coder 30B A3B3USD 0.122USD 0.206USD 0.2632.1×SiliconFlow
Qwen3.8 2.4T A95B6USD 1.75USD 3.00USD 3.752.1×Featherless
DeepSeek-V4-Flash-Vision-Exp3USD 0.323USD 0.66USD 0.662.0×DeepInfra
Qwen3 32B FP85USD 0.13USD 0.225USD 0.2632.0×DeepInfra
Gemini 3.6 Flash3USD 1.50USD 1.50USD 3.002.0×Google Gemini API, Google Vertex AI
MiniMax M3.1 Flash9USD 0.263USD 0.263USD 0.5252.0×MiniMax, DigitalOcean, Featherless, Fireworks AI, GMI Cloud, Parasail, Together AI
Qwen3.7 Max5USD 1.88USD 3.75USD 3.752.0×Novita AI
Kimi K318USD 3.76USD 6.00USD 7.502.0×Chutes
Qwen3 235B A22B FP82USD 0.4USD 0.598USD 0.7962.0×GMI Cloud
Qwen3 VL 30B A3B Instruct4USD 0.237USD 0.294USD 0.4682.0×Featherless
Qwen3.6 27B7USD 0.688USD 1.04USD 1.352.0×Chutes
Qwen3.5 35B A3B7USD 0.355USD 0.63USD 0.6881.9×DeepInfra
Gemma 4 26B A4B11USD 0.138USD 0.198USD 0.2631.9×DeepInfra, Featherless
DeepSeek V4 Pro 081315USD 1.30USD 1.98USD 2.481.9×Chutes
Mistral Large 33USD 0.75USD 0.954USD 1.381.8×Amazon Bedrock
MiniMax M2.73USD 0.367USD 0.525USD 0.6561.8×GMI Cloud
Qwen3 30B A3B (fp8)3USD 0.122USD 0.122USD 0.2151.8×Cloudflare Workers AI
GLM 5.112USD 1.40USD 2.15USD 2.371.7×Chutes
Qwen3.5 9B4USD 0.113USD 0.151USD 0.191.7×DeepInfra, SiliconFlow
Llama 4 Maverick 17B 128E Instruct FP87USD 0.388USD 0.438USD 0.6491.7×GMI Cloud
GLM 510USD 0.93USD 1.55USD 1.551.7×GMI Cloud
Llama 3 Instruct (70B)2USD 2.86USD 3.78USD 4.701.6×Amazon Bedrock
Kimi K2.58USD 0.9USD 1.20USD 1.451.6×SiliconFlow
Qwen2.5-72B-Instruct3USD 0.37USD 0.378USD 0.591.6×DeepInfra
MiMo V2.6 Flash5USD 0.14USD 0.175USD 0.2191.6×GMI Cloud
Qwen3.5 122B A10B4USD 0.715USD 1.10USD 1.101.5×SiliconFlow
Claude Sonnet 55USD 4.00USD 4.00USD 6.001.5×Anthropic, DigitalOcean, Google Vertex AI
NVIDIA Nemotron 3 Nano 30B A3B3USD 0.0875USD 0.105USD 0.1311.5×Novita AI
GLM 4.78USD 0.738USD 1.00USD 1.071.5×DeepInfra
Qwen3-30B-A3B-Instruct-25072USD 0.142USD 0.174USD 0.2061.4×SiliconFlow
Qwen3.5 27B7USD 0.588USD 0.825USD 0.8451.4×Chutes
Kimi K2.7 Code12USD 1.20USD 1.65USD 1.711.4×CoreWeave Forge
Llama 3.2 1B Instruct3USD 0.0705USD 0.0705USD 0.11.4×Cloudflare Workers AI, Featherless
GPT-5.6 Sol6USD 8.00USD 8.00USD 11.251.4×OpenAI, DigitalOcean, Microsoft Azure, Perplexity
GLM 4.5 Air4USD 0.31USD 0.372USD 0.4251.4×Novita AI
NVIDIA Nemotron 3 Ultra9USD 0.912USD 1.05USD 1.251.4×CoreWeave Forge
MiMo V2.6 Pro3USD 0.408USD 0.486USD 0.5441.3×GMI Cloud
Qwen3.8 Flash5USD 0.18USD 0.23USD 0.2371.3×DeepInfra
GLM 4.66USD 0.76USD 0.956USD 1.001.3×Venice
IBM Granite 4.2 8B2USD 0.0825USD 0.095USD 0.1071.3×CoreWeave Forge
GPT OSS Safeguard 20B3USD 0.102USD 0.131USD 0.1311.3×Amazon Bedrock
Muse Glimmer 30B3USD 0.5USD 0.525USD 0.6371.3×Featherless
Hy43USD 0.625USD 0.625USD 0.7841.3×GMI Cloud, Novita AI
Gemma 3 4B3USD 0.05USD 0.0625USD 0.06251.3×Amazon Bedrock
GPT-5.6 Terra6USD 4.50USD 4.50USD 5.631.3×OpenAI, DigitalOcean, Microsoft Azure, Perplexity
Hy4 Preview6USD 1.25USD 1.25USD 1.561.3×DeepInfra, GMI Cloud, Hyperbolic, Novita AI, SiliconFlow
Muse Spark Open4USD 2.00USD 2.00USD 2.501.3×Baseten, Fireworks AI, Together AI
Qwen3.7 Plus2USD 0.56USD 0.63USD 0.71.3×Together AI
Qwen3.5 397B A17B7USD 1.09USD 1.29USD 1.351.2×Chutes, DeepInfra
Ling 3.0 Flash VL2USD 0.09USD 0.101USD 0.1111.2×DeepInfra
Qwen3-VL 8B2USD 0.305USD 0.34USD 0.3751.2×SiliconFlow
Qwen3-14B3USD 0.122USD 0.15USD 0.151.2×SiliconFlow
Qwen3.8 Max6USD 2.48USD 3.00USD 3.001.2×DeepInfra
Qwen3.8 Max (0902)3USD 2.48USD 3.00USD 3.001.2×GMI Cloud
NVIDIA Nemotron 3.5 Lightning 30B A3B5USD 0.085USD 0.0875USD 0.1021.2×DeepInfra
Qwen3 Next 80B A3B Instruct5USD 0.343USD 0.381USD 0.4121.2×DeepInfra
gemma-4-31B-it-turbo2USD 0.152USD 0.168USD 0.1821.2×DeepInfra
Kimi K2 Thinking6USD 0.9USD 1.07USD 1.071.2×GMI Cloud
MiniMax M2.56USD 0.44USD 0.525USD 0.5251.2×Venice
Qwen3 Next 80B A3B Thinking2USD 0.412USD 0.45USD 0.4871.2×Google Vertex AI
L3.1-70B-Euryale-v2.22USD 0.72USD 0.785USD 0.851.2×Featherless
DeepSeek R1 Distill Llama 8B2USD 0.203USD 0.216USD 0.231.1×GMI Cloud
Claude Fable 55USD 20.00USD 20.00USD 22.001.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Fable 5.14USD 20.00USD 20.00USD 22.001.1×Anthropic, DigitalOcean, Google Vertex AI
Claude Haiku 4.55USD 2.00USD 2.00USD 2.201.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Opus 4.54USD 10.00USD 10.00USD 11.001.1×Anthropic, DigitalOcean, Google Vertex AI
Claude Opus 4.64USD 10.00USD 10.00USD 11.001.1×Anthropic, DigitalOcean, Google Vertex AI
Claude Opus 4.75USD 10.00USD 10.00USD 11.001.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Opus 4.85USD 10.00USD 10.00USD 11.001.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Opus 55USD 10.00USD 10.00USD 11.001.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Opus 5.56USD 8.00USD 8.00USD 8.801.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI, Perplexity
Claude Sonnet 4.54USD 6.00USD 6.00USD 6.601.1×Anthropic, DigitalOcean, Google Vertex AI
Claude Sonnet 4.65USD 6.00USD 6.00USD 6.601.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI
Claude Sonnet 5.56USD 4.00USD 4.00USD 4.401.1×Anthropic, DeepInfra, DigitalOcean, Google Vertex AI, Perplexity
GPT-5.44USD 5.63USD 5.63USD 6.191.1×OpenAI, DigitalOcean, Microsoft Azure
GPT-5.56USD 11.25USD 11.25USD 12.381.1×OpenAI, DigitalOcean, GMI Cloud, Microsoft Azure, Perplexity
GPT-6 Astra6USD 20.00USD 20.00USD 22.001.1×OpenAI, DigitalOcean, GMI Cloud, Microsoft Azure, Perplexity
GPT-6 Luna6USD 0.2USD 0.2USD 0.221.1×OpenAI, DigitalOcean, GMI Cloud, Microsoft Azure, Perplexity
GPT-6 Sol6USD 4.00USD 4.00USD 4.401.1×OpenAI, DigitalOcean, GMI Cloud, Microsoft Azure, Perplexity
GPT-6.1 Sol6USD 4.00USD 4.00USD 4.401.1×OpenAI, DigitalOcean, GMI Cloud, Microsoft Azure, Perplexity
GPT-6.1 Terra4USD 2.00USD 2.00USD 2.201.1×OpenAI, DigitalOcean, Microsoft Azure
Grok 4.66USD 3.00USD 3.00USD 3.301.1×xAI, GMI Cloud, Google Vertex AI, Microsoft Azure, Perplexity
Grok 4.76USD 3.00USD 3.00USD 3.301.1×xAI, GMI Cloud, Google Vertex AI, Microsoft Azure, Perplexity
Qwen2.5 Coder 32B Instruct2USD 0.745USD 0.777USD 0.811.1×Cloudflare Workers AI
Qwen3 235B A22B Thinking 2507 FP83USD 1.21USD 1.21USD 1.311.1×Featherless, Venice
MiniMax M24USD 0.486USD 0.525USD 0.5251.1×Featherless
Kimi K2 Instruct2USD 1.00USD 1.04USD 1.071.1×Novita AI
Inkling 2 Small3USD 0.451USD 0.475USD 0.4751.1×DeepInfra
Kimi K3.1 Pro13USD 3.80USD 4.00USD 4.001.1×DeepInfra
GLM 4.7 Flash6USD 0.145USD 0.152USD 0.1521.1×Cloudflare Workers AI
DeepSeek-V3.1-Terminus3USD 0.453USD 0.453USD 0.4641.0×GMI Cloud, SiliconFlow
Beam 1.0 501B A23B3USD 0.6USD 0.6USD 0.61.0×Baseten, DigitalOcean, Together AI
Claude Fable 5.24USD 16.00USD 16.00USD 16.001.0×Anthropic, DigitalOcean, Google Vertex AI, Perplexity
Claude Haiku 5.54USD 0.2USD 0.2USD 0.21.0×Anthropic, DigitalOcean, Google Vertex AI, Perplexity
DeepSeek-V3.2-Exp2USD 0.305USD 0.305USD 0.3051.0×GMI Cloud, SiliconFlow
Gemini 2.5 Flash3USD 0.85USD 0.85USD 0.851.0×Google Gemini API, DeepInfra, Google Vertex AI
Gemini 2.5 Flash-Lite2USD 0.175USD 0.175USD 0.1751.0×Google Gemini API, Google Vertex AI
Gemini 2.5 Pro3USD 3.44USD 3.44USD 3.441.0×Google Gemini API, DeepInfra, Google Vertex AI
Gemini 3.1 Flash-Lite (gemini-3.1-flash-lite)3USD 0.563USD 0.563USD 0.5631.0×Google Gemini API, DeepInfra, Google Vertex AI
Gemini 3.1 Pro Preview3USD 4.50USD 4.50USD 4.501.0×Google Gemini API, GMI Cloud, Google Vertex AI
Gemini 3.5 Flash3USD 3.38USD 3.38USD 3.381.0×DeepInfra, GMI Cloud, Google Vertex AI
Gemini 3.5 Flash-Lite3USD 0.85USD 0.85USD 0.851.0×Google Gemini API, GMI Cloud, Google Vertex AI
Gemini 3.7 Flash3USD 1.50USD 1.50USD 1.501.0×DeepInfra, GMI Cloud, Google Vertex AI
Gemini 3.8 Flash3USD 1.50USD 1.50USD 1.501.0×Google Gemini API, GMI Cloud, Google Vertex AI
Gemini 4 Argon4USD 4.00USD 4.00USD 4.001.0×Google Gemini API, GMI Cloud, Google Vertex AI, Perplexity
Gemini 4 Carbon4USD 4.00USD 4.00USD 4.001.0×Google Gemini API, GMI Cloud, Google Vertex AI, Perplexity
Gemini 4 Hydrogen3USD 3.00USD 3.00USD 3.001.0×Google Gemini API, Google Vertex AI, Perplexity
GLM 4.5V2USD 0.9USD 0.9USD 0.91.0×Z.ai, Novita AI
GLM 4.6V2USD 0.45USD 0.45USD 0.451.0×Z.ai, Novita AI
GLM 5.3 FlashX2USD 0.59USD 0.59USD 0.591.0×Z.ai, Novita AI
GLM-4.5-FP82USD 1.00USD 1.00USD 1.001.0×Z.ai, GMI Cloud
GPT Chat Latest2USD 11.25USD 11.25USD 11.251.0×OpenAI, Microsoft Azure
GPT-4.1 (2025-04-14)3USD 3.50USD 3.50USD 3.501.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-4.1 Mini2USD 0.7USD 0.7USD 0.71.0×OpenAI, Microsoft Azure
GPT-4o2USD 4.38USD 4.38USD 4.381.0×OpenAI, DigitalOcean
GPT-4o mini3USD 0.263USD 0.263USD 0.2631.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5 (2025-08-07)3USD 3.44USD 3.44USD 3.441.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5 Mini3USD 0.688USD 0.688USD 0.6881.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5 Nano3USD 0.138USD 0.138USD 0.1381.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5 Pro2USD 41.25USD 41.25USD 41.251.0×OpenAI, Microsoft Azure
GPT-5.12USD 3.44USD 3.44USD 3.441.0×OpenAI, Microsoft Azure
GPT-5.23USD 4.81USD 4.81USD 4.811.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5.2 Pro2USD 57.75USD 57.75USD 57.751.0×OpenAI, DigitalOcean
GPT-5.3 Codex3USD 4.81USD 4.81USD 4.811.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5.4 Mini3USD 1.69USD 1.69USD 1.691.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5.4 Nano3USD 0.463USD 0.463USD 0.4631.0×OpenAI, DigitalOcean, Microsoft Azure
GPT-5.4 Pro3USD 67.50USD 67.50USD 67.501.0×OpenAI, DigitalOcean, Microsoft Azure
Grok 4.20 (non-reasoning)2USD 1.56USD 1.56USD 1.561.0×xAI, Google Vertex AI
Grok 4.20 (reasoning)3USD 1.56USD 1.56USD 1.561.0×xAI, Google Vertex AI, Microsoft Azure
Grok 4.34USD 1.56USD 1.56USD 1.561.0×xAI, Amazon Bedrock, Google Vertex AI, Microsoft Azure
Grok 4.53USD 3.00USD 3.00USD 3.001.0×xAI, GMI Cloud, Perplexity
Kimi K2 Instruct 09052USD 1.07USD 1.07USD 1.071.0×Featherless, Novita AI
Ling 3.0 Flash2USD 0.09USD 0.09USD 0.091.0×DeepInfra, Novita AI
Minimax M2.13USD 0.525USD 0.525USD 0.5251.0×Amazon Bedrock, Featherless, GMI Cloud
Ministral 3 14B3USD 0.2USD 0.2USD 0.21.0×Mistral AI, Amazon Bedrock, DigitalOcean
Ministral 3 3B2USD 0.1USD 0.1USD 0.11.0×Mistral AI, Amazon Bedrock
Ministral 3 8B2USD 0.15USD 0.15USD 0.151.0×Mistral AI, Amazon Bedrock
o3 (2025-04-16)3USD 3.50USD 3.50USD 3.501.0×OpenAI, DigitalOcean, Microsoft Azure
Qwen3.6 Plus3USD 1.13USD 1.13USD 1.131.0×Alibaba Cloud, GMI Cloud, Together AI
Ripple-65B Preview2USD 0.3USD 0.3USD 0.31.0×GMI Cloud, Together AI
Step 3.7 Flash3USD 0.438USD 0.438USD 0.4381.0×Featherless, GMI Cloud, Novita AI
Step 5 Preview4USD 1.43USD 1.43USD 1.431.0×Featherless, GMI Cloud, Hyperbolic, Novita AI
Step-3.5-Flash2USD 0.15USD 0.15USD 0.151.0×Featherless, SiliconFlow

Is a model cheaper from its maker or from a host?

98 models with a maker's own price and at least two providers. Host ÷ maker is the cheapest third-party blended price divided by the maker's; under 1× the host is cheaper.

ModelMaker's priceCheapest hostHost priceHost ÷ maker
Qwen-MT PlusUSD 3.69Novita AIUSD 0.3750.10×
DeepSeek V4.1 Flash 1017USD 0.525DeepInfraUSD 0.1470.28×
Qwen3.8 27BUSD 1.13ChutesUSD 0.3420.30×
Qwen3.6 35B A3BUSD 0.844CoreWeave ForgeUSD 0.2970.35×
GLM 5.2USD 2.15DeepInfraUSD 0.8720.41×
Kimi K2.6USD 1.71ChutesUSD 0.7430.43×
GLM 5.3USD 2.15DeepInfraUSD 1.050.49×
GLM 5.3 FlashUSD 0.237GMI CloudUSD 0.1190.50×
Qwen3.7 MaxUSD 3.75Novita AIUSD 1.880.50×
Qwen3.6 27BUSD 1.35ChutesUSD 0.6880.51×
Qwen3.5 35B A3BUSD 0.688DeepInfraUSD 0.3550.52×
Qwen3.8 2.4T A95BUSD 3.00FeatherlessUSD 1.750.58×
GLM 5USD 1.55GMI CloudUSD 0.930.60×
Kimi K3USD 6.00ChutesUSD 3.760.63×
Qwen3.5 122B A10BUSD 1.10SiliconFlowUSD 0.7150.65×
GLM 5.1USD 2.15ChutesUSD 1.400.65×
DeepSeek V4 Pro 0813USD 1.98ChutesUSD 1.300.66×
Kimi K2.7 CodeUSD 1.71CoreWeave ForgeUSD 1.200.70×
Qwen3.5 27BUSD 0.825ChutesUSD 0.5880.71×
GLM 4.5 AirUSD 0.425Novita AIUSD 0.310.73×
GLM 4.7USD 1.00DeepInfraUSD 0.7380.74×
GLM 4.6USD 1.00VeniceUSD 0.760.76×
Qwen3.8 FlashUSD 0.23DeepInfraUSD 0.180.78×
MiniMax M3USD 0.525CoreWeave ForgeUSD 0.4120.79×
Qwen3.7 PlusUSD 0.7Together AIUSD 0.560.80×
Qwen3.5 397B A17BUSD 1.35ChutesUSD 1.090.81×
Qwen3.8 MaxUSD 3.00DeepInfraUSD 2.480.82×
Qwen3.8 Max (0902)USD 3.00GMI CloudUSD 2.480.82×
Kimi K3.1 ProUSD 4.00DeepInfraUSD 3.800.95×
Claude Fable 5USD 20.00DeepInfraUSD 20.001.00×
Claude Fable 5.1USD 20.00DigitalOceanUSD 20.001.00×
Claude Fable 5.2USD 16.00DigitalOceanUSD 16.001.00×
Claude Haiku 4.5USD 2.00DeepInfraUSD 2.001.00×
Claude Haiku 5.5USD 0.2DigitalOceanUSD 0.21.00×
Claude Opus 4.5USD 10.00DigitalOceanUSD 10.001.00×
Claude Opus 4.6USD 10.00DigitalOceanUSD 10.001.00×
Claude Opus 4.7USD 10.00DeepInfraUSD 10.001.00×
Claude Opus 4.8USD 10.00DeepInfraUSD 10.001.00×
Claude Opus 5USD 10.00DeepInfraUSD 10.001.00×
Claude Opus 5.5USD 8.00DeepInfraUSD 8.001.00×
Claude Sonnet 4.5USD 6.00DigitalOceanUSD 6.001.00×
Claude Sonnet 4.6USD 6.00DeepInfraUSD 6.001.00×
Claude Sonnet 5USD 4.00DigitalOceanUSD 4.001.00×
Claude Sonnet 5.5USD 4.00DeepInfraUSD 4.001.00×
Gemini 2.5 FlashUSD 0.85DeepInfraUSD 0.851.00×
Gemini 2.5 Flash-LiteUSD 0.175Google Vertex AIUSD 0.1751.00×
Gemini 2.5 ProUSD 3.44DeepInfraUSD 3.441.00×
Gemini 3.1 Flash-Lite (gemini-3.1-flash-lite)USD 0.563DeepInfraUSD 0.5631.00×
Gemini 3.1 Pro PreviewUSD 4.50GMI CloudUSD 4.501.00×
Gemini 3.5 Flash-LiteUSD 0.85GMI CloudUSD 0.851.00×
Gemini 3.6 FlashUSD 1.50Google Vertex AIUSD 1.501.00×
Gemini 3.8 FlashUSD 1.50GMI CloudUSD 1.501.00×
Gemini 4 ArgonUSD 4.00GMI CloudUSD 4.001.00×
Gemini 4 CarbonUSD 4.00GMI CloudUSD 4.001.00×
Gemini 4 HydrogenUSD 3.00Google Vertex AIUSD 3.001.00×
GLM 4.5VUSD 0.9Novita AIUSD 0.91.00×
GLM 4.6VUSD 0.45Novita AIUSD 0.451.00×
GLM 5.3 FlashXUSD 0.59Novita AIUSD 0.591.00×
GLM-4.5-FP8USD 1.00GMI CloudUSD 1.001.00×
GPT Chat LatestUSD 11.25Microsoft AzureUSD 11.251.00×
GPT-4.1 (2025-04-14)USD 3.50DigitalOceanUSD 3.501.00×
GPT-4.1 MiniUSD 0.7Microsoft AzureUSD 0.71.00×
GPT-4oUSD 4.38DigitalOceanUSD 4.381.00×
GPT-4o miniUSD 0.263DigitalOceanUSD 0.2631.00×
GPT-5 (2025-08-07)USD 3.44DigitalOceanUSD 3.441.00×
GPT-5 MiniUSD 0.688DigitalOceanUSD 0.6881.00×
GPT-5 NanoUSD 0.138DigitalOceanUSD 0.1381.00×
GPT-5 ProUSD 41.25Microsoft AzureUSD 41.251.00×
GPT-5.1USD 3.44Microsoft AzureUSD 3.441.00×
GPT-5.2USD 4.81DigitalOceanUSD 4.811.00×
GPT-5.2 ProUSD 57.75DigitalOceanUSD 57.751.00×
GPT-5.3 CodexUSD 4.81DigitalOceanUSD 4.811.00×
GPT-5.4USD 5.63DigitalOceanUSD 5.631.00×
GPT-5.4 MiniUSD 1.69DigitalOceanUSD 1.691.00×
GPT-5.4 NanoUSD 0.463DigitalOceanUSD 0.4631.00×
GPT-5.4 ProUSD 67.50DigitalOceanUSD 67.501.00×
GPT-5.5USD 11.25DigitalOceanUSD 11.251.00×
GPT-5.6 LunaUSD 0.45DigitalOceanUSD 0.451.00×
GPT-5.6 SolUSD 8.00DigitalOceanUSD 8.001.00×
GPT-5.6 TerraUSD 4.50DigitalOceanUSD 4.501.00×
GPT-6 AstraUSD 20.00DigitalOceanUSD 20.001.00×
GPT-6 LunaUSD 0.2DigitalOceanUSD 0.21.00×
GPT-6 SolUSD 4.00DigitalOceanUSD 4.001.00×
GPT-6.1 SolUSD 4.00DigitalOceanUSD 4.001.00×
GPT-6.1 TerraUSD 2.00DigitalOceanUSD 2.001.00×
Grok 4.20 (non-reasoning)USD 1.56Google Vertex AIUSD 1.561.00×
Grok 4.20 (reasoning)USD 1.56Google Vertex AIUSD 1.561.00×
Grok 4.3USD 1.56Amazon BedrockUSD 1.561.00×
Grok 4.5USD 3.00GMI CloudUSD 3.001.00×
Grok 4.6USD 3.00GMI CloudUSD 3.001.00×
Grok 4.7USD 3.00GMI CloudUSD 3.001.00×
MiniMax M3.1 FlashUSD 0.263DigitalOceanUSD 0.2631.00×
Ministral 3 14BUSD 0.2Amazon BedrockUSD 0.21.00×
Ministral 3 3BUSD 0.1Amazon BedrockUSD 0.11.00×
Ministral 3 8BUSD 0.15Amazon BedrockUSD 0.151.00×
o3 (2025-04-16)USD 3.50DigitalOceanUSD 3.501.00×
Qwen3.6 PlusUSD 1.13GMI CloudUSD 1.131.00×
Command RUSD 0.263Amazon BedrockUSD 0.752.86×

Which provider is cheapest most often?

Number of models, of the 197 with paid standard prices from two or more providers, where the provider has the lowest blended price. Ties count for every tied provider.

ProviderCheapest forWithout a tieModels pricedPricesWith cached price
DeepInfra49289218350%
GMI Cloud4518979763.9%
DigitalOcean410599594.9%
Google Vertex AI3636311879.4%
Microsoft Azure2909712460.8%
OpenAI2703710786.5%
Featherless198878759.8%
Perplexity1802424100%
SiliconFlow1711565648.2%
Novita AI156676865.2%
Anthropic1501841100%
Amazon Bedrock11610814631.5%

Widest coverage: Amazon Bedrock (108 models), Microsoft Azure (97 models), GMI Cloud (97 models), DeepInfra (92 models), Featherless (87 models). The full provider list is in the JSON download.

How much do batch, flex and priority tiers change the price?

Each non-standard price compared with the same provider's standard price for the same model, blended. Under 1× is a discount.

TierPricesProvidersMedian vs standard
Batch15790.50×
Flex or spot11780.54×
Priority10691.75×
Fast70152.00×
Ultrafast326.00×

How much more do output tokens cost than input tokens?

Median output price ÷ median input price, for the 424 models with a non-zero input price. Median 4.0×; middle half 3.0× to 5.0×.

Output ÷ inputModelsShare
1× or less174%
Over 1× to 2×4410.4%
Over 2× to 3×7016.5%
Over 3× to 4×11727.6%
Over 4× to 5×8720.5%
Over 5× to 8×7417.5%
Over 8×153.5%

How many providers publish a cached-input price?

65.3% of standard prices (791 of 1,212) include a price for cached input tokens. Where they do, the median cached price is 88% below the normal input price. A published cache price does not tell you the hit rate you will get; see cached input tokens.

Method

  • Source: each provider's own pricing page or price documentation, read with AI assistance and a page renderer between 2026-11-14 and 2026-11-15. Every price is stored with its source URL, the date checked and a verbatim quote from the page (src/data/own-prices/ in the site's code). No router, reseller or aggregator feed is used, not even as a cross-check.
  • Scope: 41 providers (16 of them model makers selling their own API), 1671 prices for 428 text and chat models. Image, audio and embedding models are out of scope. BroadNet Inference, which publishes this site, is not in the index.
  • Prices are USD per 1M tokens: the standard on-demand price at the lowest prompt-size tier in the provider's default region, as the page shows it. A provider's page that calls a price promotional or discounted is kept as published and counted (49 standard prices). DeepSeek bills peak and off-peak rates; the index uses its peak price.
  • Service tiers: batch, flex or spot, priority, fast and ultrafast prices are stored as separate rows (453 in all) and kept out of every per-model statistic; the service-tier table compares each with the same provider's standard price. Free rows (0 input and 0 output, 6) are also left out of price statistics.
  • Blended price = (3 × input + 1 × output) ÷ 4, a common 3:1 input-to-output mix. A model's price level is the median blended price across its paid standard prices.
  • Spread = the dearest paid standard price's blended price ÷ the cheapest's, per model, over the 197 models with paid standard prices from 2 or more providers.
  • Cheapest provider: the provider(s) with the lowest blended standard price for a model, counted over the same 197 models. Ties credit every tied provider.
  • Maker versus host: a maker's own price is a row from the model's maker selling it on its own API (for example DeepSeek on DeepSeek's API, not DeepSeek on Alibaba Cloud). Host ÷ maker compares the cheapest third-party standard price with the maker's, on the blend.
  • Cached input: a standard price that includes a cache-read price. Context tiers use the largest context any provider states for the model; models with no stated context are left out of that table.

Update frequency: recomputed on every site build from the committed provider price files. A weekly job re-fetches every source page and flags prices or quotes it no longer finds; flagged pages are re-read before any value changes. This version: checks from 2026-11-14 to 2026-11-15.

Limits: published list prices, not what any customer paid; no latency, throughput or uptime weighting; one region and the lowest prompt-size tier per provider; providers that publish no token price are absent. Code: src/lib/dataset-prices.ts.

What do the CSV columns mean?

ColumnMeaning
modelModel slug, as in the /llm-api-pricing/<model>/ page address.
nameModel name.
makerModel maker.
contextLargest context window any provider states, tokens; empty if none states it.
context_tierContext bucket used in the tier table.
providersProviders with a standard-tier price (free rows included).
standard_pricesPaid standard-tier prices used in the statistics.
median_input_per_mMedian input price across standard prices, USD per 1M tokens.
median_output_per_mMedian output price across standard prices, USD per 1M tokens.
output_input_ratiomedian_output ÷ median_input; empty if input is free.
min_blendedLowest blended standard price.
median_blendedMedian blended standard price.
max_blendedHighest blended standard price.
spreadmax_blended ÷ min_blended.
cheapest_providersProvider(s) at min_blended, separated by "; ".
priciest_providersProvider(s) at max_blended.
maker_input_per_mMaker's own API input price; empty if the maker sells none.
maker_output_per_mMaker's own API output price.
maker_blendedMaker's own blended price.
cheapest_hostCheapest third-party (non-maker) standard price, provider name.
host_vs_makerCheapest host blended ÷ maker blended; under 1 means a host is cheaper.

The JSON adds every individual price (provider, tier, input, cached, output, context, precision, promo and free flags, source URL and check date) and the provider table. For the same rows with quotes, notes and provider model IDs as a flat CSV, use the provider price database.

How to cite this dataset

Our compilation (the selection, statistics, rankings, tables and method): CC BY 4.0. Free to use, share and adapt with attribution and a link to this page.

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. Credit the provider as the source of each price.

Plain text

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/

BibTeX

@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},
}

Questions

What is the average price of an LLM API per million tokens?

The median model costs USD 0.658 per 1M tokens with input and output mixed 3 to 1, across 424 models priced on 41 providers' own pages. Half of models fall between USD 0.263 and USD 2.85.

Why does the same model have different prices?

Providers run the same open-weight model on different hardware, regions, precision and service tiers, and set their own prices. Across 197 models with two or more providers, the median gap between the cheapest and dearest standard price is 1.3×.

Is it cheaper to buy a model from its maker or from a host?

More often from the maker: a third-party host is cheaper than the maker's own API on 29 of 98 models where both publish a price, and the median cheapest host costs 1.0× the maker's price. Each model page shows the gap for that model.

Where do the prices come from?

From each provider's own pricing page, read with AI assistance and a page renderer, with the source URL, check date and a verbatim quote stored for every price. No router, reseller or aggregator feed is used. BroadNet Inference, which publishes this site, is not in the index; its prices are on the pricing page.

How often is the index updated?

When the provider price files change. A weekly job re-fetches every source page and flags any price or quote it can no longer find, for re-reading; the page, CSV, JSON and key findings are then recomputed, and each price keeps its own check date.

Can I use this data?

Yes. Our compilation (the selection, statistics, rankings, tables and method) is CC BY 4.0; cite it as shown under "How to cite". Each price is a fact published by the provider named in its row, at the source URL given; credit that provider as the source of the price.

Read as MarkdownAsk ChatGPTAsk ClaudeAsk Perplexity

Sources