Lowest published plan price for 18 AI infrastructure and inference platforms, read directly from each company's live pricing page. Every page in this table is re-fetched once a day, so these are not figures someone typed up months ago.
| Tool | Lowest plan | Line on their pricing page | Changes recorded |
|---|---|---|---|
| Together AI | $5 | $5.00 |
0 |
| Weights & Biases | $5 | $5/mo |
0 |
| Baseten | $6 | $6.60Try |
0 |
| Fireworks AI | $6 | $6.00 |
0 |
| Lambda Labs | $6.69 | $6.69 |
0 |
| Hugging Face | $9 | $9 /month |
0 |
| Pinecone | $20 | $20/month flat |
0 |
| Modal | $30 | $30 / month |
0 |
| Weaviate | $30 | $30 / organization · monthly plan |
0 |
| LangChain | $39 | $39 / seat |
0 |
| AssemblyAI | usage-based | no subscription tier published | 0 |
| Deepgram | usage-based | no subscription tier published | 0 |
| Fal.ai | usage-based | no subscription tier published | 0 |
| Groq | not detected | - | 0 |
| Qdrant | not detected | - | 0 |
| Replicate | usage-based | no subscription tier published | 0 |
| RunPod | usage-based | no subscription tier published | 0 |
| Vast.ai | usage-based | no subscription tier published | 1 |
Lowest plan = the smallest recurring plan figure currently on the pricing page ($5-$10,000 range, so metered rates, add-ons and enterprise contract figures are excluded). 6 of 18 price purely by usage and publish no subscription tier. 2 publish no figure this method can read. Plans differ in what they include; check the linked page before deciding.
What changed in the last 30 days
- Vast.ai (2026-08-15): pricing structure changed
How this is compiled
Each pricing page is fetched daily. Only price-bearing lines are extracted, normalized and hashed, so layout tweaks and copy tests do not register as price changes. When the hash moves, the diff is recorded and dated. This table is regenerated from that dataset, not maintained by hand.
Disclosure: some links in this post may be affiliate links. Pricing data is collected automatically and is not influenced by any partnership.
