THE COST OF AI

Model prices

Compare advertised text-token rates and see what your workload could cost. Explore each model for price history and a shareable chart.

Last successful collection: Sat, 17 Oct 2026 04:18:57 GMT · USD · daily snapshots

Estimate your text workload

Token charges only, in USD. Assumes uncached standard text pricing; excludes taxes, credits, tools, images, audio, routing changes and volume/context tiers. No requests are sent to these models.

63 matching models

Advertised OpenRouter rates. Prices per million tokens.
ModelInputOutputContextYour workload
Meta: Llama 3.1 8B Instruct

meta-llama

$0.05$0.08131,072$0.09
Qwen: Qwen3.7 Flash

qwen

$0.03$0.131,000,000$0.095
Meta: Llama 3.2 1B Instruct

meta-llama

$0.027$0.20160,000$0.1275
DeepSeek: DeepSeek V4 Flash 0731

deepseek

$0.07$0.141,048,576$0.14
Mistral: Ministral 3 3B 2512

mistralai

$0.10$0.10131,072$0.15
DeepSeek: DeepSeek V4 Flash 0423

deepseek

$0.091$0.1821,048,576$0.182
Meta: Llama 3.2 3B Instruct

meta-llama

$0.05$0.33131,072$0.215
Mistral: Ministral 3 8B 2512

mistralai

$0.15$0.15262,144$0.225
Meta: Llama 4 Scout

meta-llama

$0.10$0.301,310,720$0.25
Meta: Llama 3.3 70B Instruct

meta-llama

$0.10$0.32131,072$0.26
Meta: Llama Guard 4 12B

meta-llama

$0.18$0.18163,840$0.27
Mistral: Ministral 3 14B 2512

mistralai

$0.20$0.20262,144$0.30
Anthropic: Claude Haiku 5.5

anthropic · additional fees listed

$0.10$0.501,000,000$0.35
OpenAI: GPT-6 Luna

openai · additional fees listed

$0.10$0.501,050,000$0.35
OpenAI: GPT-6 Luna Pro

openai · additional fees listed

$0.10$0.501,050,000$0.35
Qwen: Qwen3.8 Flash

qwen

$0.15$0.471,000,000$0.385
Qwen: Qwen3.8 Omni Flash

qwen

$0.15$0.471,000,000$0.385
DeepSeek: DeepSeek V4.1 Flash

deepseek

$0.04$0.701,048,576$0.39
DeepSeek: DeepSeek V4 Pro 0423

deepseek

$0.2088$0.41761,048,576$0.4176
Mistral: Mistral Small 4

mistralai

$0.15$0.60262,144$0.45
DeepSeek: DeepSeek V3.2 Exp

deepseek

$0.27$0.41163,840$0.475
DeepSeek: DeepSeek V3.2

deepseek

$0.28$0.42163,840$0.49
Meta: Llama 4 Maverick

meta-llama

$0.1875$0.65251,048,576$0.5138
DeepSeek: DeepSeek V4 Flash Vision Exp

deepseek

$0.2156$0.64681,048,576$0.539
Meta: Llama 3.1 70B Instruct

meta-llama

$0.40$0.40131,072$0.60
Qwen: Qwen3.7 Plus

qwen

$0.32$1.281,000,000$0.96
Google: Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image)

google · additional fees listed

$0.25$1.5065,536$1.00
Mistral: Mistral Large 3 2512

mistralai

$0.50$1.50262,144$1.25
Qwen: Qwen3.8 27B

qwen

$0.24$2.201,000,000$1.34
Mistral: Devstral 2 2512

mistralai

$0.40$2.00262,144$1.40
Google: Gemini 3.5 Flash Lite

google · additional fees listed

$0.30$2.501,048,576$1.55
DeepSeek: DeepSeek V4 Pro 0813

deepseek

$0.66$1.981,048,576$1.65
Mistral: Mistral Large 4

mistralai

$0.68$2.091,048,576$1.725
Google: Nano Banana 2 (Gemini 3.1 Flash Image)

google · additional fees listed

$0.50$3.00131,072$2.00
SpaceXAI: Grok Build 0.1

x-ai · additional fees listed

$1.00$2.00256,000$2.00
SpaceXAI: Grok 4.20

x-ai · additional fees listed

$1.25$2.502,000,000$2.50
SpaceXAI: Grok 4.20 Multi-Agent

x-ai · additional fees listed

$1.25$2.502,000,000$2.50
SpaceXAI: Grok 4.3

x-ai · additional fees listed

$1.25$2.501,000,000$2.50
Google: Gemini 3.6 Flash

google · additional fees listed

$0.75$3.751,048,576$2.625
Google: Gemini 3.7 Flash

google · additional fees listed

$0.75$3.751,048,576$2.625
Google: Gemini 3.8 Flash

google · additional fees listed

$0.825$4.1251,048,576$2.8875
Qwen: Qwen3.8 2.4T A95B

qwen

$2.00$6.001,048,576$5.00
Qwen: Qwen3.8 Max (0902)

qwen

$2.00$6.001,000,000$5.00
SpaceXAI: Grok 4.5

x-ai · additional fees listed

$2.00$6.00500,000$5.00
SpaceXAI: Grok 4.6

x-ai · additional fees listed

$2.00$6.00500,000$5.00
SpaceXAI: Grok 4.7

x-ai · additional fees listed

$2.00$6.00500,000$5.00
Google: Nano Banana 2.1

google · additional fees listed

$1.50$7.5065,536$5.25
Mistral: Mistral Medium 3.5

mistralai

$1.50$7.50262,144$5.25
Anthropic: Claude Sonnet 5

anthropic · additional fees listed

$2.00$10.001,000,000$7.00
Anthropic: Claude Sonnet 5.5

anthropic · additional fees listed

$2.00$10.001,000,000$7.00
Google: Gemini 4 Argon

google · additional fees listed

$2.00$10.001,048,576$7.00
OpenAI: GPT-6 Sol

openai · additional fees listed

$2.00$10.001,050,000$7.00
OpenAI: GPT-6 Sol Pro

openai · additional fees listed

$2.00$10.001,050,000$7.00
OpenAI: GPT-6.1 Sol

openai · additional fees listed

$2.00$10.001,050,000$7.00
OpenAI: GPT-6.1 Sol Pro

openai · additional fees listed

$2.00$10.001,050,000$7.00
Qwen: Qwen3.8 Max Prime

qwen

$4.00$12.001,000,000$10.00
Anthropic: Claude Opus 5.5

anthropic · additional fees listed

$4.00$20.001,000,000$14.00
Anthropic: Claude Opus 4.8

anthropic · additional fees listed

$5.00$25.001,000,000$17.50
Anthropic: Claude Opus 5

anthropic · additional fees listed

$5.00$25.001,000,000$17.50
Anthropic: Claude Fable 5

anthropic · additional fees listed

$10.00$50.001,000,000$35.00
Anthropic: Claude Fable 5.1

anthropic · additional fees listed

$10.00$50.001,000,000$35.00
OpenAI: GPT-6 Astra

openai · additional fees listed

$10.00$50.001,050,000$35.00
OpenAI: GPT-6 Astra Pro

openai · additional fees listed

$10.00$50.001,050,000$35.00

What this comparison covers

Up to eight recently listed text-output models from each of eight developers: OpenAI, Anthropic, Google, DeepSeek, xAI, Qwen, Meta and Mistral. Batch-priced variants are excluded. Catalog listing dates are not verified model release dates. This is a selected catalog, not every available model.

Rates come from the public OpenRouter catalog, not direct vendor price sheets. Listed zero-price models may have access restrictions or limits. Advertised token rates do not measure quality, speed or availability. History begins when we first observe a model; missing days are never filled with invented prices.

Download prices and history as JSON