One change, in full
Drops Build/Scale/Enterprise Tiers, Adds Self-Service GPU Pricing
observed between 2025-02-08 and 2025-09-15
Before — captured 2025-02-08
Plan tiers Build/Scale/Enterprise named plans listed (free self-service Build tier with $1 credit; Scale/Enterprise via Contact sales)
HIPAA mention Listed under Scale plan
Primary page CTA "Get Started" -> self-service signup (api.together.ai)
GPU cluster self-service pricing "Starting at" rates only, no explicit self-service hourly table shown
Dedicated endpoint price (H100 80GB) $0.098/minute (~$5.88/hr equivalent)
Fine-tuning pricing transparency Interactive cost calculator only, no published per-model rate table
New product categories Not listed
After — captured 2025-09-15
Plan tiers No named plan tiers; page organized by product line with usage-based pricing only
HIPAA mention Not listed anywhere on page
Primary page CTA "Contact Sales" -> /contact-sales
GPU cluster self-service pricing New "Instant Clusters" self-service hourly table (e.g. HGX H100 SXM $2.99/hr)
Dedicated endpoint price (H100 80GB) $3.36/hour
Fine-tuning pricing transparency Explicit priced tables by model size and by named frontier model (e.g. gpt-oss-120B, Llama 4 Maverick, DeepSeek-R1)
New product categories Code Execution (Sandbox, Interpreter), Transcription, Frontier AI Factory, Storage listed with pricing
observed 2025-09-15 pricing page packagingcommercial motionprice from an archived copy on the record since 2026-10-04