Cloudflare: Models and Prices

Workers AI: serverless inference on GPUs in Cloudflare’s network (NYSE: NET).

Cloudflare served 114B tokens of other labs' open-weight models over Oct 3 to Oct 9, 2026 through a large public LLM router, #36 among providers.

Open-weight models

ModelFamilyShare of modelTokens, 7dInput $/MOutput $/MTokens/sUptime 3dContext
Gemma 4 26B A4B Gemma15.3%40B$0.10$0.304599.7%256K
DeepSeek V4 Flash 0423DeepSeek0.4%12B$0.44$1.326099.4%384K
GLM 4.7 FlashGLM100.0%8B$0.06$0.4047100.0%131K
DeepSeek V4 Pro 0423DeepSeek1.1%7B$1.15$2.555199.5%1.05M
DeepSeek V4 Flash 0731DeepSeek<0.1%5B$0.44$1.325099.6%1.05M
GLM 5.3GLM<0.1%2B$1.40$4.403598.8%1.05M
Granite 4.0 MicroOther open-weight100.0%478M$0.017$0.11236100.0%131K
GLM 5.2GLM<0.1%423M$0.50$6.008298.4%262K
Kimi K2.6Kimi0.2%224M$0.95$4.003098.9%262K
Llama 3.1 8B InstructLlama0.8%189M$0.152$0.2872199.8%32K
DeepSeek V4 Pro 0813DeepSeek--$1.32$3.9654100.0%1.05M
Qwen3.8 27BQwen--$0.45$3.204193.9%262K
Kimi K2.7 CodeKimi--$0.95$4.0036100.0%262K
Llama 3.3 70B InstructLlama--$0.293$2.255199.8%24K
Llama 3.2 3B InstructLlama--$0.051$0.335231100.0%80K
Mistral Small 3.1 24BMistral--$0.351$0.55537100.0%128K
Llama 3.2 1B InstructLlama--$0.027$0.20112199.8%60K
Qwen2.5 Coder 32B InstructQwen--$0.66$1.0032100.0%33K

List prices per million tokens as of October 9, 2026. Share: of all tokens served on that model; a dash where the volume is not reported.

Company

Headquarters
San Francisco, CA

Sources: Cloudflare Workers AI.

Finance your GPUs

Serving tokens on your own GPUs? Tell us the cluster, the models you serve and the contracts behind them, and we will share it with partner funders that fit.

Prefer email? hello@amcompute.com