Vendors — Official API Pricing
Grouped by vendor: how many models we cover, what each costs on the vendor’s own price list — plus official pricing-page links, cache-hit rates and peak/off-peak pricing that routed-price listings do not show.
OpenAI7 models
from $0.05/M in$0.4/M out
GPT-5 Nano
Anthropic4 models
from $0.1/M in$0.5/M out
Claude Haiku 5.5
cache-hit rate
Google DeepMind2 models
from $0.75/M in$3.75/M out
Gemini 3.6 Flash
cache-hit rate
Meta2 models
from $0.2/M in$0.2/M out
Meta Muse Wonder
Zhipu AI2 models
from $0.15/M in$0.5/M out
GLM-5.3-Flash
Moonshot AI2 models
from $0.95/M in$4/M out
Kimi K2.7 Code
cache-hit rate
DeepSeek2 models
from $0.15/M in$0.6/M out
DeepSeek-V4.1-Flash
peak/off-peakcache-hit rate
Alibaba Cloud / Qwen2 models
from $0.15/M in$0.47/M out
Qwen3.8-Flash (Tongyi)
Google1 models
from $2/M in$12/M out
Gemini 3.0 Pro
Mistral AI1 models
from $0.68/M in$2.09/M out
Mistral Large 4
NVIDIA1 models
from $0.5/M in$2.20/M out
Nemotron 3 Ultra
SpaceXAI (formerly xAI)1 models
from $2/M in$6/M out
Grok 4.5
ByteDance / Volcano Engine1 models
from ¥3.20/M in¥16/M out
Doubao-Seed-2.0-Pro（豆包2.0 Pro）
Xiaomi1 models
from $0.14/M in$0.28/M out
MiMo-V2.6-Flash
cache-hit rate
Baidu1 models
from ¥0.8/M in¥3.20/M out
ERNIE 4.5 Turbo
cache-hit rate
MiniMax1 models
from $0.3/M in$1.20/M out
MiniMax M3
cache-hit rate
Tencent1 models
from $0.834/M in$2.50/M out
腾讯混元 Hy4 preview
cache-hit rate
StepFun1 models
from $1/M in$2.70/M out
Step 5 Preview
cache-hit rate