Cheapest AI model for each task

Ranked from the live catalog: coding, chat, translation, OCR, images, video, embeddings and 3D. Provider list price, zero platform markup — the same numbers the price feed publishes.

Cheapest for coding agents

Ranked by one 1M-input + 200K-output run. Prices are provider list price with 0% platform markup.

ModelProviderCost
glm-4.1v-thinking-flashZhiPuFREE
glm-4.6v-flashZhiPuFREE
glm-4.7-flashZhiPuFREE
hy-mt2-liteTencent Hunyuan$0.077 per run
deepseek-v4-flashDeepSeek$0.114 per run

Cheapest for general chat

Ranked by one 1M-input + 200K-output run. Prices are provider list price with 0% platform markup.

ModelProviderCost
glm-4.1v-thinking-flashZhiPuFREE
glm-4.6v-flashZhiPuFREE
glm-4.7-flashZhiPuFREE
hy-mt2-liteTencent Hunyuan$0.077 per run
deepseek-v4-flashDeepSeek$0.114 per run

Cheapest for translation

Ranked by per 1M input tokens. Prices are provider list price with 0% platform markup.

ModelProviderCost
hy-mt2-liteTencent Hunyuan$0.043 per 1M input
hy-mt2-plusTencent Hunyuan$0.072 per 1M input
qwen-mt-plusAlibaba Cloud$0.250 per 1M input
hy-mt2-proTencent Hunyuan$0.291 per 1M input

Cheapest for OCR / vision extraction

Ranked by per 1M input tokens. Prices are provider list price with 0% platform markup.

ModelProviderCost
qwen-vl-ocrAlibaba Cloud$0.700 per 1M input
qwen3.5-ocrAlibaba Cloud$0.700 per 1M input

Cheapest per image

Ranked by one generated image. Prices are provider list price with 0% platform markup.

ModelProviderCost
cogview-3-flashZhiPuFREE
wan2.2-t2i-flashAlibaba Cloud$0.0215
doubao-seedream-5.0ByteDance$0.0270
wan2.5-t2i-previewAlibaba Cloud$0.0308
doubao-seedream-5.0-proByteDance$0.0340

Cheapest per 5-second video

Ranked by one 5s clip (estimated). Prices are provider list price with 0% platform markup.

ModelProviderCost
pixverse-video-c1PixVerse$0.154
pixverse-video-v5.6PixVerse$0.203
pixverse-video-v6.0PixVerse$0.203
hy-video-v1.5Tencent Hunyuan$0.231
doubao-seedance-2.0-fastByteDance$0.270

Cheapest embeddings

Ranked by per 1M input tokens. Prices are provider list price with 0% platform markup.

ModelProviderCost
kinfra-text-embedding-0.6bTencent$0.077
kinfra-vl-embedding-2bTencent$0.077
text-embedding-v4Alibaba Cloud$0.080
kinfra-text-embedding-4bTencent$0.092
kinfra-vl-embedding-8bTencent$0.092

Cheapest 3D generation

Ranked by per generation. Prices are provider list price with 0% platform markup.

ModelProviderCost
hy-3d-motionTencent Hunyuan$0.185
hy-3d-expressTencent Hunyuan$0.277
tripo-3d-3.1Tripo$0.323
tripo-3d-p1Tripo$0.323
hy-3d-3.0Tencent Hunyuan$0.369
How this is calculated. Text models are ranked by the cost of one run of 1M input + 200K output tokens; video by one 5-second clip (token-billed models converted at their rate cards); images per generated image; embeddings per 1M input tokens; 3D per generation.
Raw data: china-ai-leaderboard.json · Verification: transparency · Free tier: 5 free models

Not sure what a model id means? Decode it here.