A compact vision-language model under Apache 2.0, easy to host next to your KYC data.
A compact dense vision-language model under Apache 2.0.
Input: text, image, video
Harnesses in Runtime
Underwriting
For KYC and KYB checks, our top pick is Qwen3.8 27B from Qwen (Alibaba). A compact vision-language model under Apache 2.0, easy to host next to your KYC data. GLM-5.3 Flash is a close second, and DeepSeek V4.1 Flash rounds out the list.
KYC and KYB agents check identity documents, business filings, and beneficial owners, then flag mismatches for review. The inputs are images and PII, so vision and the ability to self-host matter most.
Updated
A compact vision-language model under Apache 2.0, easy to host next to your KYC data.
A compact dense vision-language model under Apache 2.0.
Input: text, image, video
Harnesses in Runtime
Image input, MIT license, and low cost for high onboarding volume.
A natively multimodal GLM that Z.ai says beats GLM-5.2 at one-tenth the price.
Input: text, image
Harnesses in Runtime
Fast multimodal checks with an MIT license when you need throughput.
A multimodal MoE that activates only 8B parameters per token and shrinks the KV cache about 4x.
Input: text, image
Harnesses in Runtime
Turn the steps your team already follows into a skill the agent reads before every run.
Run two or three of these models against past cases. Evals show which one passes at the lowest cost.
The agent works on its own computer in your VPC, with role-based access and an approval before sensitive actions.
For KYC and KYB checks, our top pick is Qwen3.8 27B from Qwen (Alibaba). A compact vision-language model under Apache 2.0, easy to host next to your KYC data. GLM-5.3 Flash is a close second, and DeepSeek V4.1 Flash rounds out the list.
Image input for IDs and filings. Self-hostable, so PII stays in your environment. Permissive license. Test two or three candidates against your own past cases before you commit, and keep a human approval on any step that moves money or changes a customer's status.
Yes. Download the weights and serve them with vLLM or SGLang in your own cloud, or use Amazon Bedrock or Google Vertex AI in your account. With Runtime, the agent's computer also runs in your VPC, so card data and PII stay in your environment.
Qwen3.8 27B (Apache 2.0), GLM-5.3 Flash (MIT), DeepSeek V4.1 Flash (MIT) are permissive and fine for commercial use.