[metadata]
description: DeepInfra startup program. Get access to up to 50B tokens for free inference. Apply now! With DeepInfra, it's easy to deploy the latest state-of-the-art AI models in production
hostname: https://deepinfra.com
keywords: inference,ml,ai,whisper,stable,diffusion
og:description: DeepInfra startup program. Get access to up to 50B tokens for free inference. Apply now! With DeepInfra, it's easy to deploy the latest state-of-the-art AI models in production
og:image: https://deepinfra.com/open-graph.png
og:image:type: image/png
og:locale: en_US
og:site_name: DeepInfra
og:title: DeepStart | Production-Ready Machine Learning Models | DeepInfra
og:url: https://deepinfra.com/deepstart
theme-color: #2A3275
twitter:card: summary_large_image
twitter:description: DeepInfra startup program. Get access to up to 50B tokens for free inference. Apply now! With DeepInfra, it's easy to deploy the latest state-of-the-art AI models in production
twitter:image: https://deepinfra.com/open-graph.png
twitter:site: @DeepInfra
twitter:title: DeepStart | Production-Ready Machine Learning Models | DeepInfra
viewport: initial-scale=1, width=device-width

[canonical-links]
https://deepinfra.com/deepstart

[document-links]
/
/Claude: /claude
/DeepSeek: /deepseek
/Flux: /flux
/Gemini: /gemini
/Kimi: /kimi
/Llama: /llama
/Mistral: /mistral
/Nemotron: /nemotron
/Qwen: /qwen
About: /about
Automatic Speech Recognition: /models/automatic-speech-recognition
Blog: /blog
Bria / remove_background_guided: /Bria/remove_background_guided
Careers: /careers
Chat: /chat
Compare: /compare
Contact Sales: /contact-sales
Contact us: /contact-sales
DeepCluster: /deepcluster
DeepGPT: https://deepgpt.com
DeepStart: /deepstart
Docs: https://docs.deepinfra.com
Embeddings: /models/embeddings
GPUs: /gpu-instances
Log In: /login
Media Center: /media-center
Models: /models
Pricing: /pricing
Privacy Policy: /privacy
Qwen / Qwen3.6-35B-A3B: /Qwen/Qwen3.6-35B-A3B
Reranker: /models/reranker
Status: https://status.deepinfra.com
Terms of Service: /terms
Text Generation: /models/text-generation
Text To Image: /models/text-to-image
Text To Music: /models/text-to-music
Text To Speech: /models/text-to-speech
Text To Video: /models/text-to-video
Trust Center: https://trust.deepinfra.com
View all models: /models
View full collection (100+): /models
World Model: /models/world-model
XiaomiMiMo / MiMo-V2.6-Pro: /XiaomiMiMo/MiMo-V2.6-Pro
Zero Shot Image Classification: /models/zero-shot-image-classification
anthropic / claude-sonnet-5-5: /anthropic/claude-sonnet-5-5
black-forest-labs / FLUX-2-klein-4b: /black-forest-labs/FLUX-2-klein-4b
deepseek-ai / DeepSeek-V4-Pro: /deepseek-ai/DeepSeek-V4-Pro
google / embeddinggemma-2: /google/embeddinggemma-2
google / gemma-4-31B-it: /google/gemma-4-31B-it
https://discord.gg/x88dCvhqYq
https://github.com/DeepInfra
https://linkedin.com/company/deep-infra
https://x.com/DeepInfra
inclusionAI / Ling-3.1-flash: /inclusionAI/Ling-3.1-flash
read the announcement: /series-b
tencent / Hy4-preview: /tencent/Hy4-preview
text-generation XiaomiMiMo/ MiMo-V2.6-Flash MiMo-V2.6-Flash is the efficiency-balanced model in Xiaomi's MiMo-V2.6 series: a sparse Mixture-of-Experts model with 309B total and 15B activated parameters, natively omnimodal across text, image, video and audio, with a 1M-token context window. Trained with large-scale mixed reinforcement learning across coding, general agents, visual and cybersecurity tasks, it is built for agentic workloads such as long-horizon coding, tool use and computer use. ZDR fp8 1024k $0.003 cached, $0.14 in, $0.28 out / 1M: /XiaomiMiMo/MiMo-V2.6-Flash
text-generation XiaomiMiMo/ MiMo-V2.6-Pro MiMo-V2.6-Pro is the flagship of Xiaomi's MiMo-V2.6 series: a sparse Mixture-of-Experts model with 1.02T total and 42B activated parameters, natively omnimodal across text, image, video and audio, with a 1M-token context window. Trained with large-scale mixed reinforcement learning and groupwise agentic grading, it is built for demanding agentic work including long-horizon coding, multi-tool agents and cybersecurity tasks. ZDR fp8 1024k $0.004 cached, $0.43 in, $0.87 out / 1M: /XiaomiMiMo/MiMo-V2.6-Pro
text-generation deepseek-ai/ DeepSeek-V4-Flash DeepSeek V4 Flash is an efficiency-focused MoE model with 284B total parameters (13B active) and a 1M-token context window. It's tuned for fast inference and high-throughput use cases while still holding up on reasoning and coding tasks. ZDR Priority Flex Cache retention fp8 1024k $0.018 cached, $0.09 in, $0.18 out / 1M: /deepseek-ai/DeepSeek-V4-Flash
text-generation deepseek-ai/ DeepSeek-V4-Flash-0731 DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities. DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on benchmarks listed below despite its far smaller activated parameter count, and is broadly competitive with the strongest proprietary models available. ZDR Priority Flex fp8 1024k $0.015 cached, $0.06 in, $0.18 out / 1M: /deepseek-ai/DeepSeek-V4-Flash-0731
text-generation deepseek-ai/ DeepSeek-V4-Pro-0813 DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro, superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in production environments. It is built on the DeepSeek-V4-Pro (Preview) model structure, with a DSpark speculative decoding module attached. ZDR Priority Flex fp8 1024k $0.10 cached, $1.30 in, $2.60 out / 1M: /deepseek-ai/DeepSeek-V4-Pro-0813
text-generation deepseek-ai/ DeepSeek-V4.1-Flash Introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. ZDR Priority Flex fp8 1024k 30% off $0.004 cached, $0.14 in, $0.42 out / 1M: /deepseek-ai/DeepSeek-V4.1-Flash
text-generation moonshotai/ Kimi-K3 Kimi K3 is Moonshot AI's 2.8T-parameter open-weight multimodal reasoning model. Built for complex coding, knowledge work, and long-horizon agentic workflows, it excels at navigating large repositories, calling tools, debugging, and iterating on images, logs, tests, and runtime feedback. Its architecture relies on KDA and Attention Residuals for computational efficiency. ZDR Cache retention mxfp4 1024k $0.285 cached, $2.85 in, $14.25 out / 1M: /moonshotai/Kimi-K3
text-generation nvidia/ NVIDIA-Nemotron-3-Ultra-550B-A55B Nemotron 3 Ultra is built for, frontier reasoning, orchestration, coding agents, deep research, and complex enterprise workflows. It delivers up to 5x faster inference and up to 30% lower cost for agentic workloads while supporting up to 1M token context. ZDR Priority Flex Cache retention fp4 256k $0.10 cached, $0.50 in, $2.20 out / 1M: /nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B
text-generation tencent/ Hy4-preview Hy4 preview is a 770B-parameter Mixture-of-Experts flagship model from the Tencent Hy Team, with 49B parameters activated per token and a 1M-token context window. It uses Gated DeepSeek Sparse Attention with IndexCache, and is post-trained for software engineering, office and data analysis, game development, and scientific research. ZDR fp8 1024k $0.042 cached, $0.834 in, $2.501 out / 1M: /tencent/Hy4-preview
text-generation zai-org/ GLM-5.2 GLM-5.2 is Z-AI's latest flagship model for long-horizon tasks. It marks a substantial leap in long-horizon task capability over its predecessor GLM-5.1 and, for the first time, delivers that capability on a **solid 1M-token context**. ZDR Priority Flex Cache retention fp4 1024k 25% off $0.105 cached, $0.563 in, $1.80 out / 1M: /zai-org/GLM-5.2
text-generation zai-org/ GLM-5.3 GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves on GLM-5.2 in coding and in the balance between performance and token efficiency. ZDR Flex fp4 1024k 38% off $0.125 cached, $0.563 in, $2.50 out / 1M: /zai-org/GLM-5.3
text-generation zai-org/ GLM-5.3-Flash GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while reducing compute overhead. ZDR fp4 1024k 50% off $0.015 cached, $0.075 in, $0.25 out / 1M: /zai-org/GLM-5.3-Flash

[page-chrome]
Models
By Category
Automatic Speech Recognition
Embeddings
Reranker
Text Generation
Text To Image
Text To Music
Text To Speech
Text To Video
World Model
Zero Shot Image Classification
View all models
By Family
/Claude
/DeepSeek
/Flux
/Gemini
/Kimi
/Llama
/Mistral
/Nemotron
/Qwen
Docs
Pricing
GPUs
Chat
DeepStart
DeepCluster
Blog
Contact Sales
Log In

[content]
DeepStart | Production-Ready Machine Learning Models | DeepInfra
DeepInfra raises $107M Series B to scale the inference cloud —
read the announcement
Models
Automatic Speech Recognition
Embeddings
Reranker
Text Generation
Text To Image
Text To Music
Text To Speech
Text To Video
World Model
Zero Shot Image Classification
Docs
Pricing
GPUs
Chat
DeepStart
DeepCluster
Blog
Feedback
Contact Sales
Log In
DeepStart
~50,000,000,000
free tokens*
To qualify your startup needs to meet the following
Have raised between 250K and 10M USD
Founded in the last 2 years
* at DeepSeek V4 Flash prices
Credits expire 1 year after approval
Apply now
Featured models
What we loved, used, and implemented most last month
View All
text-generation
text-to-speech
text-to-image
text-generation
tencent/
Hy4-preview
Hy4 preview is a 770B-parameter Mixture-of-Experts flagship model from the Tencent Hy Team, with 49B parameters activated per token and a 1M-token context window. It uses Gated DeepSeek Sparse Attention with IndexCache, and is post-trained for software engineering, office and data analysis, game development, and scientific research.
ZDR
fp8
1024k
$0.042 cached, $0.834 in, $2.501 out / 1M
text-generation
XiaomiMiMo/
MiMo-V2.6-Flash
MiMo-V2.6-Flash is the efficiency-balanced model in Xiaomi's MiMo-V2.6 series: a sparse Mixture-of-Experts model with 309B total and 15B activated parameters, natively omnimodal across text, image, video and audio, with a 1M-token context window. Trained with large-scale mixed reinforcement learning across coding, general agents, visual and cybersecurity tasks, it is built for agentic workloads such as long-horizon coding, tool use and computer use.
ZDR
fp8
1024k
$0.003 cached, $0.14 in, $0.28 out / 1M
text-generation
XiaomiMiMo/
MiMo-V2.6-Pro
MiMo-V2.6-Pro is the flagship of Xiaomi's MiMo-V2.6 series: a sparse Mixture-of-Experts model with 1.02T total and 42B activated parameters, natively omnimodal across text, image, video and audio, with a 1M-token context window. Trained with large-scale mixed reinforcement learning and groupwise agentic grading, it is built for demanding agentic work including long-horizon coding, multi-tool agents and cybersecurity tasks.
ZDR
fp8
1024k
$0.004 cached, $0.43 in, $0.87 out / 1M
text-generation
deepseek-ai/
DeepSeek-V4.1-Flash
Introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens.
ZDR
Priority
Flex
fp8
1024k
30% off
$0.004 cached, $0.14 in, $0.42 out / 1M
text-generation
zai-org/
GLM-5.3
GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks. It supports text input and output with a 1M-token context window, and improves on GLM-5.2 in coding and in the balance between performance and token efficiency.
ZDR
Flex
fp4
1024k
38% off
$0.125 cached, $0.563 in, $2.50 out / 1M
text-generation
zai-org/
GLM-5.3-Flash
GLM-5.3-Flash is a native multimodal model from Z.ai. It is suited for efficient coding and long-horizon agent tasks. Its hybrid sparse and linear attention architecture maintains accurate long-context behavior while reducing compute overhead.
ZDR
fp4
1024k
50% off
$0.015 cached, $0.075 in, $0.25 out / 1M
text-generation
moonshotai/
Kimi-K3
Kimi K3 is Moonshot AI's 2.8T-parameter open-weight multimodal reasoning model. Built for complex coding, knowledge work, and long-horizon agentic workflows, it excels at navigating large repositories, calling tools, debugging, and iterating on images, logs, tests, and runtime feedback. Its architecture relies on KDA and Attention Residuals for computational efficiency.
ZDR
Cache retention
mxfp4
1024k
$0.285 cached, $2.85 in, $14.25 out / 1M
text-generation
deepseek-ai/
DeepSeek-V4-Pro-0813
DeepSeek-V4-Pro-0813 is the official release of DeepSeek-V4-Pro, superseding the preview version, with greatly enhanced agentic capabilities and performance improvements that are especially pronounced in production environments. It is built on the DeepSeek-V4-Pro (Preview) model structure, with a DSpark speculative decoding module attached.
ZDR
Priority
Flex
fp8
1024k
$0.10 cached, $1.30 in, $2.60 out / 1M
text-generation
deepseek-ai/
DeepSeek-V4-Flash-0731
DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities. DeepSeek-V4-Flash-0731 outperforms DeepSeek-V4-Pro (Preview) on benchmarks listed below despite its far smaller activated parameter count, and is broadly competitive with the strongest proprietary models available.
ZDR
Priority
Flex
fp8
1024k
$0.015 cached, $0.06 in, $0.18 out / 1M
text-generation
zai-org/
GLM-5.2
GLM-5.2 is Z-AI's latest flagship model for long-horizon tasks. It marks a substantial leap in long-horizon task capability over its predecessor GLM-5.1 and, for the first time, delivers that capability on a **solid 1M-token context**.
ZDR
Priority
Flex
Cache retention
fp4
1024k
25% off
$0.105 cached, $0.563 in, $1.80 out / 1M
text-generation
nvidia/
NVIDIA-Nemotron-3-Ultra-550B-A55B
Nemotron 3 Ultra is built for, frontier reasoning, orchestration, coding agents, deep research, and complex enterprise workflows. It delivers up to 5x faster inference and up to 30% lower cost for agentic workloads while supporting up to 1M token context.
ZDR
Priority
Flex
Cache retention
fp4
256k
$0.10 cached, $0.50 in, $2.20 out / 1M
text-generation
deepseek-ai/
DeepSeek-V4-Flash
DeepSeek V4 Flash is an efficiency-focused MoE model with 284B total parameters (13B active) and a 1M-token context window. It's tuned for fast inference and high-throughput use cases while still holding up on reasoning and coding tasks.
ZDR
Priority
Flex
Cache retention
fp8
1024k
$0.018 cached, $0.09 in, $0.18 out / 1M
View full collection (100+)
Have questions or need a custom solution?
Contact Sales
Company
Pricing
Docs
Compare
DeepStart
About
Careers
Contact us
Media Center
Trust Center
Status
DeepGPT
Latest Models
Bria
/
remove_background_guided
inclusionAI
/
Ling-3.1-flash
google
/
embeddinggemma-2
anthropic
/
claude-sonnet-5-5
tencent
/
Hy4-preview
Featured Models
deepseek-ai
/
DeepSeek-V4-Pro
google
/
gemma-4-31B-it
XiaomiMiMo
/
MiMo-V2.6-Pro
Qwen
/
Qwen3.6-35B-A3B
black-forest-labs
/
FLUX-2-klein-4b
© 2026 DeepInfra. All rights reserved.
Privacy Policy
Terms of Service
