Hyperbolic resets GPU rates upward — H100 SXM to $2.89/hr
Hyperbolic nearly doubled its published H100 SXM starting rate to $2.89/GPU/hr, raised H200 to $3.49 and B200 to $5.99, and pulled its consumer-GPU per-hour rates.
H100 SXM $1.50/GPU/hr, H200 $2.40, B200 $3.50, RTX 4090 $0.30, RTX 3070 $0.16 — five GPU types with published per-hour starting rates.
H100 SXM $2.89/GPU/hr, H200 $3.49, B200 $5.99 — three GPU types published, with the wider catalog advertised only as from $0.20/GPU/hr.
Proof of change
Hyperbolic's own wording, on the page where we captured it.
pricing is refreshed weekly based on the best available rates from suppliers on our platform
Quoted from the capture below — these words appear on the page as shown, not paraphrased.
Showing the whole page as captured — scroll the panel, or open it at full size.
This is a single observation, not a before/after comparison — the change it describes has no visible transition to photograph, so we show the page that states it instead. The pixels are our own capture, unmodified.
Hyperbolic’s on-demand GPU page now publishes only three starting rates, all materially higher than the June 2026 capture: NVIDIA H100 SXM at $2.89 / HR (up 93%), H200 at $3.49 / HR (up 45%), and B200 at $5.99 / HR (up 71%). The per-hour rates previously shown for RTX 4090 and RTX 3070 are gone; the page instead advertises the catalog as starting at $0.20/GPU/hr. The note that “pricing is refreshed weekly based on the best available rates from suppliers on our platform” is unchanged, so the move reflects Hyperbolic’s aggregated third-party supply repricing rather than a new list price.
Serverless inference is untouched — the per-million-token rate card still runs $0.10 for Llama-3.2-3B / Llama-3.1-8B, $0.20 for Qwen2.5-Coder-32B, $0.40 for the 70B-class SKUs, $2.00 for DeepSeek-V2.5 and $4.00 for Llama-3.1-405B, with Basic (60 RPM), Pro (600 RPM) and Enterprise (unlimited) tiers gating request rate rather than price.
Published on-demand starting rates move to H100 SXM $2.89, H200 $3.49 and B200 $5.99 per GPU-hour (catalog advertised from $0.20/GPU/hr), and the consumer-GPU per-hour rates are pulled from the page. Packaging is now four surfaces — On-Demand, Reserved, Private Cloud and Serverless Inference — funded by prepaid compute credits (minimum $5, never expiring) with Auto Top-Up, a per-instance price lock, and 99.5%/99.9% uptime SLAs. Serverless token rates are unchanged.