Vercel
Adobe
Ubisoft
Iconic Hearts
Latitude
Abacus.AI company wordmark
Hugging Face smiling emoji logo with wordmark
interface.ai company wordmark
Salesforce cloud logo
Requesty company wordmark

Funding news · May 4, 2026

Deepinfra closes a $107M Series B to grow the inference cloud

The round was led jointly by 500 Global and Georges Harik, with NVIDIA, Felicis, Samsung Next, Supermicro, Crescent Cove and several other backers joining. Deepinfra is putting it toward capacity and new hardware.

Start building Good timing for a launch.

$107M

Series B raised

25x

more tokens served since Series A

Trillions of tokens, on a budget that holds

Usage-based billing with no lock-in and no surprise line items. Two founders in a garage or a global enterprise — Deepinfra grows with you, through clean APIs and engineers who actually answer.

Inference shaped around your workload

Tune for latency, throughput, cost or sheer volume. Deepinfra builds the deployment around whichever matters most to you, with more than 100 hosted models ready to call.

Nothing retained. Fully audited. Locked down.

Prompts, completions and end-user data are never stored. Deepinfra is certified for ISO 27001 and SOC 2, and treats security reviews as routine work rather than a yearly scramble.

Our own GPUs. Our own racks. Your speed advantage.

Deepinfra operates its own GPU fleet inside secured US facilities, tuned end to end for serving models — so latency stays low and uptime stays high.

Models

Browse featured Deepinfra models

Speech and image models sit in the full Deepinfra catalogue. Open it here

Moonshot AI logomoonshotai/
text-generation

Deepinfra catalogue cover tile: white square with the Moonshot AI swirl logo above small grey Kimi-K3.1-Pro text

Open-weight multimodal agent model tuned with large-scale RL and a 1M-token window, served on Deepinfra.

Agentic · multimodal
1024kTrending
ZDRPriorityFlex
DeepSeek whale logodeepseek-ai/
text-generation

White square model cover with the blue DeepSeek whale logo and small V4.1-Flash-1017 caption

Incremental refresh of V4.1-Flash with sharper reasoning, coding and agent scores, plus shorter answers.

Reasoning · code
1024kTrending
ZDRPriorityFlexfp8
NVIDIA eye logonvidia/
text-generation

White square cover showing the green NVIDIA eye logo with a small Nemotron 3.5 Super VL label

120B mixture-of-experts vision model (12B active) that reads images, documents, screens and video, then acts.

Vision · agents
ZDRfp4256k
tencent/
text-generation

White square cover with the blue Tencent logo above the words Hy4-preview in blue

Early look at Tencent's 770B-parameter mixture-of-experts flagship, callable through the Deepinfra API.

MoE · preview
ZDRfp41024k
Xiaomi MiMo logoXiaomiMiMo/
text-generation

Dark navy square cover with a glowing blue-purple MiMo M logo and V2.6 Flash caption

Efficiency-focused sparse MoE from Xiaomi's V2.6 line, built for quick, lightweight everyday calls.

Speed · chat
ZDRfp81024k
Xiaomi MiMo logoXiaomiMiMo/
text-generation

Dark navy square cover with a glowing blue-purple MiMo M logo and V2.6 Pro caption

Top model of Xiaomi's V2.6 family: a 1.02T-parameter sparse MoE aimed at hard reasoning work.

Deep reasoning
Trending
ZDRfp81024k
DeepSeek whale logodeepseek-ai/
text-generation

White square cover with a faint blue DeepSeek whale and DeepSeek-V4.1-Flash caption

Multimodal MoE on a 552B backbone, balancing speed and depth for agents and chat on Deepinfra.

Multimodal · MoE
1024kTrending
ZDRPriorityFlexfp8
Z.ai logozai-org/
text-generation

White square cover with a black Z.ai logo tile and small GLM-5.3 label

Z.ai's large reasoning model for complex software engineering and long-horizon agent tasks.

Coding · agents
Trending
ZDRFlexfp41024k
Z.ai logozai-org/
text-generation

White square cover with a black Z.ai logo tile and small GLM-5.3-Flash label

Natively multimodal sibling of GLM-5.3, sized for quick coding help and long agent runs.

Coding · fast
ZDRfp41024k
Moonshot AI logomoonshotai/
text-generation

White square cover with the dark Moonshot AI swirl logo and small Kimi-K3 caption

Moonshot's 2.8T-parameter open-weight multimodal reasoner for demanding coding jobs, hosted on Deepinfra.

Reasoning · code
mxfp41024k
ZDRCache retention
DeepSeek whale logodeepseek-ai/
text-generation

White square cover with soft blue swooshes and small DeepSeek-V4-Pro-0813 text

Official V4-Pro release replacing the preview build, with broad gains in reasoning quality.

Reasoning
1024k
ZDRPriorityFlexfp8
DeepSeek whale logodeepseek-ai/
text-generation

White square cover with the blue DeepSeek whale and small V4-Flash-0731 caption

Official V4-Flash release that supersedes the preview, keeping latency low for high-volume traffic.

High volume
1024k
ZDRPriorityFlexfp8

Inference metrics, streaming live

Speed, scale, stability and spend across the Deepinfra fleet, refreshed as traffic flows.

40.42M

Tokens / second

345ms

Time to first token

3654

Requests / second

24.35

exaFLOPS online

DeepCluster

A private NVIDIA B300 cluster, operated by Deepinfra

Hardware sourced, racked and run by Deepinfra in a Tier 3 facility. You own the capacity; we keep it healthy under a 99.982% uptime SLA.

Plan your cluster

NVIDIA B300 · 5-year term

99.982%

uptime, written into the SLA


288 GB
HBM3e memory / GPU
256–5,000
GPUs per cluster
Tier 3
data center
Get started
Get started