Baseten Starter Credits ($30) + Startup Program (up to $25K)
Source: https://www.baseten.co/pricing
Description
Create account to comment on specific lines or Sign in
+ 1 Baseten is a managed AI inference / model deployment platform (autoscaling GPU pods, OpenAI-compatible Model APIs, Chains for multi-step inference, Training, cross-cloud routing). Every new workspace gets $30 in free starter credits automatically — enough to deploy a small open-source model and benchmark it for a few hours. Early-stage AI startups can layer on the AI Startup Program: up to $25,000 for Dedicated Inference / Training and up to $2,500 for Model APIs. Past mass promo "Fall into Inference" (up to $10K for new customers) ended 2025-12-20 and is no longer available — the $30 + Startup Program is the current standing offer as of May 2026.
No comments on this line yet.
+ 2
No comments on this line yet.
+
3
No comments on this line yet.
+ 4
No comments on this line yet.
+ 6
No comments on this line yet.
+ 7 1. Go to baseten.co and click "Get started" (top-right) or sign up directly at app.baseten.co/signup.
No comments on this line yet.
+ 8 2. Sign up with email/password, Google, or GitHub.
No comments on this line yet.
+ 9 3. Verify your email — Baseten sends a confirmation link.
No comments on this line yet.
+ 10 4. Create a workspace (name + region). The workspace is the billing/team boundary.
No comments on this line yet.
+ 11 5. No credit card required to receive the $30 starter credits — they are credited automatically to the new workspace on creation.
No comments on this line yet.
+ 12 6. (Optional) Add a payment method only if you want to keep running deployments after the $30 is exhausted.
No comments on this line yet.
+ 13 7. Go to Model library or Deploy from Hugging Face / Truss to spin up your first model.
No comments on this line yet.
+ 14
No comments on this line yet.
+ 15 Important:
No comments on this line yet.
+ 16 • $30 is a one-time, per-workspace grant. Spinning up a second workspace under the same email/org will not stack credits.
No comments on this line yet.
+ 17 • Credits cover all usage on the Startup plan: GPU compute time (pods), Model API calls, Training runs, cold-start, scaling-down time.
No comments on this line yet.
+ 18 • You only pay for time the model is actively deploying, scaling up/down, or making predictions — scale-to-zero is supported (no charge while idle).
No comments on this line yet.
+ 19 • There is no monthly platform fee on the Startup plan — pure usage-based.
No comments on this line yet.
+ 20
No comments on this line yet.
+
21
No comments on this line yet.
+ 22
No comments on this line yet.
+ 24
No comments on this line yet.
+ 25 GPU pricing (May 2026, on-demand per-minute billing — see pricing page):
No comments on this line yet.
+ 26
No comments on this line yet.
+ 27 GPUApprox. $/hourHours from $30
No comments on this line yet.
+ 28 T4 (16 GB)~$0.63~47 hours
No comments on this line yet.
+ 29 A10G (24 GB)~$1.21~25 hours
No comments on this line yet.
+ 30 L4 (24 GB)~$1.10~27 hours
No comments on this line yet.
+ 31 A100 40 GB~$3.93~7.6 hours
No comments on this line yet.
+ 32 A100 80 GB~$5.12~5.8 hours
No comments on this line yet.
+ 33 H100 (80 GB)~$9.98~3 hours
No comments on this line yet.
+ 34 H100 MIG (40 GB)~$5.49~5.4 hours
No comments on this line yet.
+ 35 B200 (180 GB)~$10.94~2.7 hours
No comments on this line yet.
+ 36
No comments on this line yet.
+ 37 For Model APIs (managed endpoints for popular OSS models — DeepSeek V3, Llama 3.3 70B, Qwen3, embeddings, etc.), $30 is enough for millions of input tokens on smaller models, or hundreds of thousands of tokens on flagship 70B+ models. Exact $/1M token rates vary by model and are listed on the pricing page.
No comments on this line yet.
+ 38
No comments on this line yet.
+
39
No comments on this line yet.
+ 40
No comments on this line yet.
+ 42
No comments on this line yet.
+ 43 If you're a venture-backed AI-first startup, this is where the real money is.
No comments on this line yet.
+ 44
No comments on this line yet.
+ 46
No comments on this line yet.
+ 47 • AI-first — AI must be core to your product (not bolted-on)
No comments on this line yet.
+ 48 • Funding stage: Seed to Series A
No comments on this line yet.
+ 49 • Age: founded less than 5 years ago
No comments on this line yet.
+ 50 • Net new to Baseten — never received Baseten credits before (the $30 starter does not disqualify you from applying)
No comments on this line yet.
+ 51
No comments on this line yet.
+ 53
No comments on this line yet.
+ 54 • Up to $25,000 in credits for Dedicated Inference or Training (GPU compute)
No comments on this line yet.
+ 55 • Up to $2,500 in credits for Model APIs (managed OSS endpoints)
No comments on this line yet.
+ 56 • Faster support: response within a few business days, dedicated Slack channel, video calls with engineers
No comments on this line yet.
+ 57 • Networking: invite-only events with Baseten engineers, founders, and other portfolio startups
No comments on this line yet.
+ 58 • GTM amplification: help with product launches, funding announcements
No comments on this line yet.
+ 59
No comments on this line yet.
+ 61
No comments on this line yet.
+ 62 1. Fill out the application form at baseten.co/talk-to-us/startup-program/
No comments on this line yet.
+ 63 2. Provide company name, funding stage, AI use case, and rough monthly inference need
No comments on this line yet.
+ 64 3. Baseten reviews and follows up — expect a few business days for response
No comments on this line yet.
+ 65 4. If approved, credits are added to your workspace and you get access to the Startup-tier perks
No comments on this line yet.
+ 66
No comments on this line yet.
+ 67 Note: Credit amounts are "up to" — the actual grant depends on your stage and projected usage. Smaller seed-stage applicants typically get less than $25K.
No comments on this line yet.
+ 68
No comments on this line yet.
+
69
No comments on this line yet.
+ 70
No comments on this line yet.
+ 72
No comments on this line yet.
+ 73 • Nebius x Baseten — Nebius credit holders can redeem free Baseten credits via Nebius's partner program. Worth checking if you have unused Nebius credits.
No comments on this line yet.
+ 74 • JPMorgan Startup Offers — Baseten is listed in JPMorgan's startup vendor catalog; JPMorgan-backed startups may get bespoke discounts/credits.
No comments on this line yet.
+ 75 • "Fall into Inference" (EXPIRED) — Was up to $10K for new customers. Ended 2025-12-20. Don't expect this to come back in the same form, but Baseten typically runs a similar Q4 promo each year — worth checking baseten.co/talk-to-us/ in autumn 2026.
No comments on this line yet.
+ 76
No comments on this line yet.
+
77
No comments on this line yet.
+ 78
No comments on this line yet.
+ 80
No comments on this line yet.
+ 82
No comments on this line yet.
+ 83 Flagship OSS models hosted by Baseten (May 2026, see model library):
No comments on this line yet.
+ 84 • DeepSeek V3 / DeepSeek R1
No comments on this line yet.
+ 85 • Llama 3.3 70B, Llama 3.1 8B / 70B / 405B
No comments on this line yet.
+ 86 • Qwen3 family (incl. Qwen3-Coder, Qwen3-VL)
No comments on this line yet.
+ 87 • Mistral Large 2, Mistral Small
No comments on this line yet.
+ 88 • Whisper Large v3 (STT)
No comments on this line yet.
+ 89 • Embedding models (BGE, E5, Jina, mixedbread)
No comments on this line yet.
+ 90 • Image: FLUX.1, SDXL
No comments on this line yet.
+ 91
No comments on this line yet.
+ 93
No comments on this line yet.
+ 94 • Deploy any Hugging Face model or custom Truss package
No comments on this line yet.
+ 95 • Multi-GPU, multi-node training and inference
No comments on this line yet.
+ 96 • Chains — multi-step inference graphs (e.g., RAG, agentic workflows)
No comments on this line yet.
+ 97 • Cross-cloud routing — Baseten can transparently route between AWS, GCP, OCI, Nebius, CoreWeave for capacity/price
No comments on this line yet.
+ 98 • Self-hosted Baseten — run the Baseten control plane in your own VPC (Enterprise only)
No comments on this line yet.
+ 99
No comments on this line yet.
+ 100 Always verify the current model list on baseten.co/library/ — new models are added weekly.
No comments on this line yet.
+ 101
No comments on this line yet.
+
102
No comments on this line yet.
+ 103
No comments on this line yet.
+ 105
No comments on this line yet.
+ 106 PlanPlatform feeBest for
No comments on this line yet.
+ 107 Startup (default for new workspaces)$0 — pure usage-basedSolo devs, indie builders, early-stage startups
No comments on this line yet.
+ 108 ProNegotiatedTeams with higher monthly volume needing custom rates
No comments on this line yet.
+ 109 EnterpriseCustom (often $5K+/mo)SLAs, self-hosted control plane, dedicated capacity, SSO/SOC2
No comments on this line yet.
+ 110
No comments on this line yet.
+ 111 The $30 starter credit applies to Startup-plan workspaces. Enterprise contracts negotiate credits separately.
No comments on this line yet.
+ 112
No comments on this line yet.
+
113
No comments on this line yet.
+ 114
No comments on this line yet.
+ 116
No comments on this line yet.
+
117
• Scale-to-zero is your friend. Configure your deployment with a low min_replicas (0 or 1) and aggressive scale_down_after so you don't burn credits while idle. Cold starts are typically 10–60s for small models, longer for 70B+.
No comments on this line yet.
+
118
• Truss > raw Docker. Baseten's Truss packaging format is the path of least resistance — wrap a Hugging Face pipeline in a Truss and it deploys with truss push.
No comments on this line yet.
+ 119 • Watch for B200/H100 cold-start cost. Even one minute of an H100 spinning up costs $0.16. Don't deploy a 70B model on H100 "to test" — try a small T4 deployment first.
No comments on this line yet.
+ 120 • Model API vs Dedicated: for low/sporadic traffic, Model APIs (per-token billing) are cheaper than keeping a dedicated GPU warm. Switch to dedicated only when you have steady throughput.
No comments on this line yet.
+ 121 • No region lock-in for credits — credits work across all Baseten-supported regions and clouds.
No comments on this line yet.
+ 122 • No credit-card hold at signup, unlike most cloud providers — this makes it a good "just try it" platform.
No comments on this line yet.
+ 123
No comments on this line yet.
+
124
No comments on this line yet.
+ 125
No comments on this line yet.
+ 126 Sources:
No comments on this line yet.
+ 127 • Baseten Pricing
No comments on this line yet.
+ 128 • Usage-based pricing with free credits — changelog
No comments on this line yet.
+ 129 • Baseten AI Startup Program
No comments on this line yet.
+ 130 • Apply: Baseten Startup Program
No comments on this line yet.
+ 131 • Fall into Inference (expired 2025-12-20)
No comments on this line yet.
+ 132 • Baseten Pricing 2026 — Costbench
No comments on this line yet.
+ 133 • Baseten on JPMorgan Startup Offers
No comments on this line yet.
+ 134 • Baseten Model Library
No comments on this line yet.
+ 135 • Truss documentation
No comments on this line yet.