# Baseten > Lowest measured time to first token, with OpenAI and Anthropic-compatible endpoints. Canonical: https://www.subconscious.dev/providers/baseten · By The Subconscious Team · Updated September 30, 2026 - Founded: 2019 - Example models: GLM 5.2, gpt-oss 120B - Website: https://www.baseten.co ## Overview Baseten runs two products. Model APIs serve a curated set of 13 open models, including DeepSeek V4, GLM 5.2, Kimi K3 and gpt-oss 120B, over endpoints that speak both the OpenAI Chat Completions shape and the Anthropic Messages shape. That dual compatibility means an existing OpenAI or Claude SDK, or a coding agent, points at Baseten with a base URL change. Dedicated deployments take any model you package with the open-source Truss CLI and bill per GPU minute, with an H100 at about $6.50 an hour. Baseten posted the lowest time to first token on the Artificial Analysis provider board in August 2026, 0.49 seconds, and its stack does KV cache-aware routing that pays off on agentic coding traffic. It also sells white-label infrastructure to model labs. Poolside launched its Laguna models on a Baseten-run branded API with metering, per-key limits and sampled traffic feeding back into training, live within 48 hours of signup. Founded in 2019, the company has raised over $585M and reportedly sits at a $13B valuation after a 2026 Series F. ## Upsides - Lowest measured time to first token among the major open-model hosts. - OpenAI and Anthropic API compatibility on every endpoint. - Per-minute GPU billing with scale to zero and a 99.99% uptime SLA. - Self-host, HIPAA and data residency options for regulated buyers. ## Core use cases - Serving private fine-tunes or custom non-LLM models like speech and embeddings. - Model labs that want a white-label production API without building one. ## Downsides - A 13-model catalog means anything off-list pushes you into running your own deployment. - Effective H100 pricing runs well above DeepInfra and Together clusters. ## At a glance | Attribute | Value | |---|---| | Model access | Open weights, 13 curated | | Flagship models | GLM 5.2, DeepSeek V4, Kimi K3, gpt-oss 120B | | Speed | 0.49s TTFT, lowest measured | | Price | H100 about $6.50/hr dedicated | | Customization | Deploy any model with Truss | | Deployment | Model APIs, dedicated, self-host | | Long context | Varies by model | ## FAQ ### What is Baseten? Baseten runs two products. Model APIs serve a curated set of 13 open models, including DeepSeek V4, GLM 5.2, Kimi K3 and gpt-oss 120B, over endpoints that speak both the OpenAI Chat Completions shape and the Anthropic Messages shape. That dual compatibility means an existing OpenAI or Claude SDK, or a coding agent, points at Baseten with a base URL change. Dedicated deployments take any model you package with the open-source Truss CLI and bill per GPU minute, with an H100 at about $6.50 an hour. ### What is Baseten best for? Serving private fine-tunes or custom non-LLM models like speech and embeddings; Model labs that want a white-label production API without building one. ### How much does Baseten cost? Baseten pricing at a glance: H100 about $6.50/hr dedicated. Rates change often, so check Baseten's pricing page before committing. ### How much context does Baseten support? Baseten's long-context support: Varies by model. ### What are the downsides of Baseten? A 13-model catalog means anything off-list pushes you into running your own deployment; Effective H100 pricing runs well above DeepInfra and Together clusters. ### What are the best alternatives to Baseten? Common alternatives include Subconscious, OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock. Each has a head-to-head comparison with Baseten on this site. ## Comparisons - [Subconscious vs Baseten](https://www.subconscious.dev/compare/subconscious-vs-baseten.md) - [OpenAI vs Baseten](https://www.subconscious.dev/compare/openai-vs-baseten.md) - [Anthropic vs Baseten](https://www.subconscious.dev/compare/anthropic-vs-baseten.md) - [Google Vertex AI vs Baseten](https://www.subconscious.dev/compare/google-vertex-vs-baseten.md) - [Amazon Bedrock vs Baseten](https://www.subconscious.dev/compare/aws-bedrock-vs-baseten.md) - [Together AI vs Baseten](https://www.subconscious.dev/compare/together-ai-vs-baseten.md) - [Fireworks AI vs Baseten](https://www.subconscious.dev/compare/fireworks-vs-baseten.md) - [Baseten vs Groq](https://www.subconscious.dev/compare/baseten-vs-groq.md) - [Baseten vs Cerebras](https://www.subconscious.dev/compare/baseten-vs-cerebras.md) - [Baseten vs DeepInfra](https://www.subconscious.dev/compare/baseten-vs-deepinfra.md) - [Baseten vs Modal](https://www.subconscious.dev/compare/baseten-vs-modal.md) - [Baseten vs xAI](https://www.subconscious.dev/compare/baseten-vs-xai.md) - [Baseten vs DeepSeek](https://www.subconscious.dev/compare/baseten-vs-deepseek.md) - [Baseten vs Moonshot AI](https://www.subconscious.dev/compare/baseten-vs-moonshot-ai.md) - [Baseten vs Z.ai](https://www.subconscious.dev/compare/baseten-vs-z-ai.md) - [Baseten vs Alibaba Cloud](https://www.subconscious.dev/compare/baseten-vs-alibaba-cloud.md) - [Baseten vs Meta](https://www.subconscious.dev/compare/baseten-vs-meta.md) - [Baseten vs SambaNova](https://www.subconscious.dev/compare/baseten-vs-sambanova.md) - [Baseten vs Nebius](https://www.subconscious.dev/compare/baseten-vs-nebius.md) - [Baseten vs fal](https://www.subconscious.dev/compare/baseten-vs-fal.md) - [Baseten vs Novita AI](https://www.subconscious.dev/compare/baseten-vs-novita-ai.md) - [Baseten vs Parasail](https://www.subconscious.dev/compare/baseten-vs-parasail.md) - [Baseten vs Inference.net](https://www.subconscious.dev/compare/baseten-vs-inference-net.md) - [Baseten vs GMI Cloud](https://www.subconscious.dev/compare/baseten-vs-gmi-cloud.md) - [Baseten vs Sail Research](https://www.subconscious.dev/compare/baseten-vs-sail-research.md) - [Baseten vs Morph](https://www.subconscious.dev/compare/baseten-vs-morph.md) - [Baseten vs Relace](https://www.subconscious.dev/compare/baseten-vs-relace.md) - [Baseten vs TypeSafe AI](https://www.subconscious.dev/compare/baseten-vs-typesafe-ai.md) - [Baseten vs StepFun](https://www.subconscious.dev/compare/baseten-vs-stepfun.md) - [Baseten vs Runware](https://www.subconscious.dev/compare/baseten-vs-runware.md) - [Baseten vs StreamLake](https://www.subconscious.dev/compare/baseten-vs-streamlake.md) - [Baseten vs Wafer](https://www.subconscious.dev/compare/baseten-vs-wafer.md) - [Baseten vs RunInfra](https://www.subconscious.dev/compare/baseten-vs-runinfra.md) - [Baseten vs Particle.AI](https://www.subconscious.dev/compare/baseten-vs-particle-ai.md) ## Sources - [Baseten review, Continuum](https://continuumcode.ai/guides/baseten-review/) - [Poolside case study](https://www.baseten.co/resources/customers/how-baseten-powered-poolsides-model-launch-in-record-time/) Pricing and model lineups change often; figures are a snapshot.