Hyperbolic
Hyperbolic provides discounted on-demand inference on open models and GPU rentals, granting a small free inference credit after phone verification.
- Free credit
- $1
- Card required
- No
Hyperbolic scores 4.9/10 for free value in LLM. The main constraint: Reported $1 inference credit granted after phone-number verification at signup; .edu users have at times gotten $15 of H100 credit. Finite, eligibility/verification gated.
Free value score
4.9 / 10Show the math
Key metrics
Models
0 free · 26 totalLimits may be stale — last re-verified outside this section's freshness window.
| Model | Family | AA | Access |
|---|---|---|---|
| GPT-OSS 120B (openai/gpt-oss-120b) | GPT | 23.8 | paid |
| DeepSeek-R1 (deepseek-ai/DeepSeek-R1) | DeepSeek | 20.1 | paid |
| Kimi K2 Instruct (moonshotai/Kimi-K2-Instruct) | Kimi | 19.4 | paid |
| GPT-OSS 20B (openai/gpt-oss-20b) | GPT | 14.9 | paid |
| DeepSeek-V3 (deepseek-ai/DeepSeek-V3) | DeepSeek | 10.4 | paid |
| Llama 3.1 405B BASE (meta-llama/Meta-Llama-3.1-405B) | Llama | 8.5 | paid |
| Meta-Llama-3.1-405B-Instruct (meta-llama/Meta-Llama-3.1-405B-Instruct) | Llama | 8.5 | paid |
| Qwen2.5-Coder-32B-Instruct (Qwen/Qwen2.5-Coder-32B-Instruct) | Qwen | 7.1 | paid |
| DeepSeek-R1-0528 (deepseek-ai/DeepSeek-R1-0528) | DeepSeek | — | paid |
| DeepSeek-V3-0324 (deepseek-ai/DeepSeek-V3-0324) | DeepSeek | — | paid |
| FLUX.1-dev (FLUX.1-dev) | FLUX | — | paid |
| Llama 3.3 70B Instruct (meta-llama/Llama-3.3-70B-Instruct) | Llama | — | paid |
| MeloTTS (MeloTTS) | MeloTTS | — | paid |
| NVIDIA Nemotron Nano 12B v2 VL (nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-BF16) | Nemotron | — | paid |
| Pixtral 12B (mistralai/Pixtral-12B-2409) | Mistral | — | paid |
| Qwen2.5-72B-Instruct (Qwen/Qwen2.5-72B-Instruct) | Qwen | — | paid |
| Qwen2.5-VL-72B-Instruct (Qwen/Qwen2.5-VL-72B-Instruct) | Qwen | — | paid |
| Qwen2.5-VL-7B-Instruct (Qwen/Qwen2.5-VL-7B-Instruct) | Qwen | — | paid |
| Qwen3-235B-A22B (Qwen/Qwen3-235B-A22B) | Qwen | — | paid |
| Qwen3-Coder 480B (Qwen/Qwen3-Coder-480B-A35B-Instruct) | Qwen | — | paid |
| Qwen3-Next 80B Instruct (Qwen/Qwen3-Next-80B-A3B-Instruct) | Qwen | — | paid |
| Qwen3-Next 80B Thinking (Qwen/Qwen3-Next-80B-A3B-Thinking) | Qwen | — | paid |
| QwQ-32B (Qwen/QwQ-32B) | Qwen | — | paid |
| SDXL Turbo (SDXL-turbo) | Stable Diffusion | — | paid |
| Stable Diffusion 2 / 1.5 (SD2, SD1.5) | Stable Diffusion | — | paid |
| Stable Diffusion XL 1.0 (SDXL1.0-base) | Stable Diffusion | — | paid |
What you get
Reported $1 inference credit after phone verification
Discounted on-demand inference on open models
Also rents GPUs (H100/H200) for self-hosting
Available as a Hugging Face inference provider
Gotchas
- $1 one-time credit
The Basic (free) tier grants a one-time $1 promotional credit that is consumed by pay-as-you-go usage and never regenerates.
- Phone verification required
The $1 promotional credit is only granted after verifying a phone number at signup.
- Personal/internal use only
The Terms of Use grant API access solely for personal or internal business purposes and prohibit reselling, sublicensing, or distributing the service.
Sources
-
"| **Basic** | 60 | 100 | $0 | API access, All models, Community support |"
https://www.hyperbolic.ai/docs/inference/overview -
"Verify & Claim → Verify your phone number to receive a $1 credit for inference services."
https://airdropalert.com/airdrops/hyperbolic-gpu-marketplace/ -
"We're giving $15 in compute credits to the first 1,000 signups. Register now and start building."
https://www.hyperbolic.ai/blog/monthly-recap-may-2025 -
"Artificial Analysis gives V3.1 a 60 score in reasoning mode, up slightly from R1's score of 59."
https://the-decoder.com/deepseeks-first-hybrid-model-v3-1-outperforms-its-r1-reasoning-model-on-benchmarks/
