Model APIs made for products, not toys
On-demand frontier models running on the Baseten Inference Stack that won’t ruin launch day.
With Baseten, we now support open-source models like DeepSeek and Llama in Retool, giving users more flexibility for what they can build. Our customers are creating AI apps and workflows, and Baseten's Model APIs deliver the enterprise-grade performance and reliability they need to ship to production.
With Baseten, we now support open-source models like DeepSeek and Llama in Retool, giving users more flexibility for what they can build. Our customers are creating AI apps and workflows, and Baseten's Model APIs deliver the enterprise-grade performance and reliability they need to ship to production.
Build your product with pre-optimized frontier models
Baseten Model APIs are built for production first, with the performance and reliability that only the Baseten Inference Stack can enable.
Ship faster
Use our Model APIs as drop-in replacements for closed models with comprehensive observability, logging, and budgeting built in.
Scale further
Run leading open-source models on our optimized infra with the fastest runtime available, all on the latest-generation GPUs.
Spend less
Spend 5-10x less than closed alternatives with our optimized multi-cloud infrastructure and efficient frontier open models.
Fast inference that scales with you
Try out new models, integrate them into your product, and launch to the top of Hacker News and Product Hunt—all in a single day.
OpenAI compatible
Migrate from closed models to open-source by swapping a URL. We’re fully OpenAI compatible with support for function calling and more.
Pre-optimized performance
We ship leading models optimized from the bottom up with the Baseten Inference Stack, so every Model API is ultra-fast out of the box.
Seamless scaling
Go from Model API to dedicated deployments on the hardware of your choosing in two clicks from the Baseten UI.
Four nines of uptime
We achieve reliability that only active-active redundancy can provide with our cloud-agnostic, multi-cluster autoscaling.
Secure and compliant
We take extensive security measures, never store inference inputs or outputs, and are SOC 2 Type II certified and HIPAA compliant.
Featureful inference
Structured outputs and tool use are baked into our Model APIs as part of the Baseten Inference Stack.
Instant access to leading models
Model library
DeepSeek V4 Pro 0813
DeepSeek V4 Pro 0813 is the latest 1.6T-parameter open frontier model from DeepSeek AI
Kimi K3
The open frontier model. 2.8T parameters, 1M-token context, top benchmarks scores.
Kimi K2.6
Kimi K2.6 builds on Kimi K2.5 with increased agentic capabilities
GLM-5.2 Fast
A speed-optimized GLM-5.2 Model API built for real-time workloads
GLM-5.2
GLM-5.2 is Z.AI's next-gen model for agentic engineering, with stronger coding and agentic capabilities and sustained execution on long-horizon tasks.
DeepSeek-V4-Flash-0731
DeepSeek V4-Flash-0731 is an open-weight 284B MoE (13B active) with 1M context and selectable reasoning effort, tuned for coding, chat, and agent workflows.
Inkling
Inkling is a multimodal MoE model (975B total, 41B active, 256k context) reasoning over text, image, and audio inputs.
Inkling-Small
An efficient, open-weight multimodal AI model. Enjoy faster inference, lower costs, and native text, image, and audio support at 1M tokens.
NVIDIA Nemotron 3 Ultra
550B hybrid Mamba-Transformer MoE with 55B active params, latent MoE routing, multi-token prediction, and up to 1M token context
Pricing
Price per
1M tokens
Model
Input
Cache Input
Output
Built for every stage in your inference journey
Explore resourcesRime's state-of-the-art p99 latency and 100% uptime is driven by our shared laser focus on fundamentals, and we're excited to push the frontier even further with Baseten.
Rime's state-of-the-art p99 latency and 100% uptime is driven by our shared laser focus on fundamentals, and we're excited to push the frontier even further with Baseten.
