Skip to content
English

Switch theme

// Models

Pick, price and run AI models

Price, performance and access in one place: what each model costs per million tokens, how smart and fast it is, and which plan, free option or own hardware fits.

models priced
263
vendors
36
official prices
165
with speed data
437

01What do you want to do?

02Price quick view and champions

Price quick view – flagship modelsAll prices →

ModelInputOutputSource
GPT-6 AstraOpenAI $10$50 Official
Claude Opus 5.5Anthropic $4$20 Official
Gemini 4 ArgonGoogle $2$10 Official
DeepSeek V4.1 Flash 1017 (deepseek-flash)DeepSeek ¥2¥8 Official
Qwen3.8-MaxAlibaba (Qwen) ¥12¥36 Official
Grok 4.7xAI $2$6 Official
GLM-5.3Zhipu (Z.ai) ¥8¥28 Official
Kimi K3Moonshot (Kimi) ¥20¥100 Official
MiniMax-M3.1-FlashMiniMax ¥1¥4 Official
Mistral Large 4Mistral AI $0.68$2.09 Official
Doubao-Seed-2.1-turboByteDance (Doubao) ¥3¥15 Official Unverified
Step-5-PreviewStepFun (阶跃星辰) ¥7¥20 Official

Per 1M tokens, original currency (¥ = CNY). Checked 2026-10-22. Full table & calculator →

Best value (AA ≥ 40)

  1. 1DeepSeek V4.1 Flash (1017)DeepSeek · AA 45.1$0.147
  2. 2Claude Haiku 5.5Anthropic · AA 43.4$0.2
  3. 3GLM-5.3-FlashZhipu (Z.ai) · AA 41.8$0.194
  4. 4MiniMax-M3.1-FlashMiniMax · AA 42.7$0.261
  5. 5MiMo-V2.6-ProXiaomi (MiMo) · AA 46.3$0.486
  6. 6Step-5-PreviewStepFun (阶跃星辰) · AA 43.7$1.53
  7. 7Gemini 3.8 FlashGoogle · AA 40.9$1.5
  8. 8Muse Spark 1.3Meta · AA 48.1$2

AA index ÷ blended USD price (official price preferred).

Fastest output

  1. 1Mercury 2.5773 tok/s
  2. 2Nemotron 3.5 Super VL429 tok/s
  3. 3Gemini 3.5 Flash-Lite404 tok/s
  4. 4Ling 3.0 Flash358 tok/s
  5. 5Beam 1.0 501B A23B327 tok/s
  6. 6Nemotron 3.5 Lightning307 tok/s

Artificial Analysis medians, not our own tests.

Cheapest official prices

  1. 1Qwen3.7-FlashAlibaba (Qwen)¥0.2/0.8
  2. 2Qwen-TurboAlibaba (Qwen)¥0.3/0.6
  3. 3Qwen-FlashAlibaba (Qwen)¥0.15/1.5
  4. 4Qwen3-VL-FlashAlibaba (Qwen)¥0.15/1.5
  5. 5GLM-4.6V-FlashXZhipu (Z.ai)¥0.15/1.5
  6. 6Qwen3.5-FlashAlibaba (Qwen)¥0.2/2

03Model library: 262 models from 36 vendors

Every model we have a price for. Open a model to see its price on every channel, context, modalities, licence, speed and benchmarks.

Alibaba (Qwen) (46)

OpenAI (35)

Anthropic (20)

Google (17)

Zhipu (Z.ai) (14)

DeepSeek (12)

Meta (11)

Mistral AI (11)

ByteDance (Doubao) (8)

Cohere (8)

Moonshot (Kimi) (7)

NVIDIA (7)

xAI (7)

Tencent (Hunyuan) (6)

Baidu (ERNIE) (5)

MiniMax (5)

Amazon (4)

Ant (inclusionAI) (4)

StepFun (阶跃星辰) (4)

Xiaomi (MiMo) (4)

Nous Research (3)

TheDrummer (3)

AI21 Labs (2)

Inception (2)

Microsoft (2)

Perplexity (2)

Sao10K (2)

ScaleDown (2)

Thinking Machines (2)

Gryphe (1)

IBM (1)

Meituan (LongCat) (1)

Moondream (1)

Reflection (1)

Upstage (1)

Writer (1)

Prices checked 2026-10-22; benchmarks updated 2026-10-22.

Data updated: Model API prices · HF model metadata(UTC+8)