Focused model portfolio
The models teams are asking to put into production now.
Six models.
Clear prices.
Enterprise hosting.
Qwen 3.8, DeepSeek V4 Flash, GLM 5.2, Kimi K3, Qwen 3.6—and MiniMax H3.
30% lowerthan current OpenRouter rates
Below OpenRouter
Public starting prices benchmarked against current listed rates.
Email to start
No funnel. Tell the founder what you want to run.
Enterprise hosting.
Transparent starting prices.
Token prices are per 1M tokens. Every listed starting price is 30% below the current OpenRouter rate used for this comparison.
qwen/qwen3.8-27b
Input$0.28 / 1M
Output$2.10 / 1M
Host this model deepseek-v4-flash-0731
Input$0.0385 / 1M
Output$0.0919 / 1M
Host this model z-ai/glm-5.2
Input$0.2352 / 1M
Output$0.7392 / 1M
Host this model moonshotai/kimi-k3
Input$1.82 / 1M
Output$9.10 / 1M
Host this model qwen/qwen3.6-35b-a3b
Input$0.049 / 1M
Output$0.49 / 1M
Host this model Pricing benchmark verified August 23, 2026 against the lowest active OpenRouter endpoint rates available for these listings. Provider prices can change. Final pricing is subject to volume, capacity, and deployment agreement.
Image + video
MiniMax H3.
Bring next-generation visual generation into your enterprise workflow, with a clear starting point for both video and image workloads.
$0.091per video second
$0.028per image
A shorter path from model to production.
Start with the workload—not a generic self-serve catalog.
Direct conversation
Email the founder with the model, traffic profile, and deployment needs. The first discussion stays focused on fit.
Transparent starting point
See the public usage prices before the conversation. No hidden calculator or gated pricing page.
Focused portfolio
Six deliberately selected models across text, image, and video instead of a sprawling, undifferentiated catalog.
Start the conversation
Tell us what you want to run.
Share your model, expected volume, and deployment requirements. We’ll reply directly.
founder@lighterhub.app