Quick request
130 open-weight models we can start on a GPU server for you. Pick one and send the request; most are running within about 30 minutes of our confirmation.
- Reserved by the hour
- The server runs your model and nobody else’s traffic. You pay per hour while it runs: what the GPUs cost us, plus a margin.
- Or per token
- Once a model carries steady traffic it moves to shared per-token prices, set so a server that is busy 50% of the time pays for itself. Where that lands well above what other providers charge, we quote it instead.
- Nothing billed before you agree
- We reply with the price and the start time first. The time shown is for downloading the weights and starting the server; the largest models also depend on GPUs being free.
Under $5 an hour
Small and mid-size models on one or two GPUs.
| Model | Size | Runs on | Ready in | Reserved | Per million tokens | Request |
|---|---|---|---|---|---|---|
| Gemma 4 26B A4Bgoogle/gemma-4-26b-a4b-it | 25.8B (4.4B active) | 1x L40S | About 15 min | $1.40/hour | $0.095 in, $0.317 out | Request |
| Llama 3.1 8B Instructmeta-llama/llama-3.1-8b-instruct | 8B | 1x L40S | About 15 min | $1.40/hour | $0.043 in, $0.081 out | Request |
| Gemma 4 31Bgoogle/gemma-4-31b-it | 31.3B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Mistral Nemomistralai/mistral-nemo | 12.2B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| gpt-oss-120bopenai/gpt-oss-120b | 117B (5.2B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.114 in, $0.569 out | Request |
| Gemma 3 4Bgoogle/gemma-3-4b-it | 4.3B (3.6B active) | 1x L40S | About 15 min | $1.40/hour | $0.021 in, $0.035 out | Request |
| gpt-oss-20bopenai/gpt-oss-20b | 20.9B (3.6B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.043 in, $0.198 out | Request |
| Qwen3.8 27Bqwen/qwen3.8-27b | 27.8B | 1x A100 | About 15 min | $2.25/hour | $0.193 in, $2.50 out | Request |
| Qwen3.5-35B-A3Bqwen/qwen3.5-35b-a3b | 36B (4.2B active) | 1x A100 | About 15 min | $2.25/hour | $0.043 in, $0.324 out | Request |
| Qwen3.6 35B A3Bqwen/qwen3.6-35b-a3b | 36B (4.2B active) | 1x A100 | About 15 min | $2.25/hour | $0.040 in, $0.372 out | Request |
| Qwen3.5-9Bqwen/qwen3.5-9b | 9.7B (8.6B active) | 1x L40S | About 15 min | $1.40/hour | $0.047 in, $0.070 out | Request |
| Gemma 3 27Bgoogle/gemma-3-27b-it | 27.4B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Qwen3 VL 8B Instructqwen/qwen3-vl-8b-instruct | 8.8B | 1x L40S | About 15 min | $1.40/hour | $0.054 in, $0.162 out | Request |
| Gemma 3 12Bgoogle/gemma-3-12b-it | 12.2B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Mistral Small 3.2 24Bmistralai/mistral-small-3.2-24b-instruct | 24B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Qwen3 32Bqwen/qwen3-32b | 32.8B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Llama 4 Scoutmeta-llama/llama-4-scout | 109B (17B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | Quoted | Request |
| Qwen3 30B A3B Instruct 2507qwen/qwen3-30b-a3b-instruct-2507 | 30.5B (3B active) | 1x L40S | About 15 min | $1.40/hour | $0.088 in, $0.293 out | Request |
| GLM 4.7 Flashz-ai/glm-4.7-flash | 31.2B (4.9B active) | 1x L40S | About 15 min | $1.40/hour | $0.054 in, $0.335 out | Request |
| Qwen3 VL 30B A3B Instructqwen/qwen3-vl-30b-a3b-instruct | 31.1B (3.6B active) | 1x L40S | About 15 min | $1.40/hour | $0.107 in, $0.372 out | Request |
| Nemotron 3.5 Lightningnvidia/nemotron-3.5-lightning | 32.9B (2.6B active) | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Muse Glimmer 30Bmeta/muse-glimmer-30b | 29.8B | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.314 in, $1.30 out | Request |
| Mistral Small 3mistralai/mistral-small-24b-instruct-2501 | 23.6B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Ministral 3 3B 2512mistralai/ministral-3b-2512 | 3.8B | 1x L40S | About 15 min | $1.40/hour | $0.029 in, $0.029 out | Request |
| Llama Guard 4 12Bmeta-llama/llama-guard-4-12b | 12B | 1x L40S | About 15 min | $1.40/hour | $0.086 in, $0.086 out | Request |
| gpt-oss-safeguard-20bopenai/gpt-oss-safeguard-20b | 21.5B (4.2B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.043 in, $0.171 out | Request |
| Mistral Small 4mistralai/mistral-small-2603 | 119B (6.5B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.163 in, $0.65 out | Request |
| Hy-MT2-30B-A3Btencent/hy-mt2-30b-a3b | 30.1B (3.2B active) | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Llama 3 8B Lunarissao10k/l3-lunaris-8b | 8B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Ministral 3 8B 2512mistralai/ministral-8b-2512 | 8.9B | 1x L40S | About 15 min | $1.40/hour | $0.094 in, $0.094 out | Request |
| Qwen3 Next 80B A3B Instructqwen/qwen3-next-80b-a3b-instruct | 81.3B (5.2B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.055 in, $0.667 out | Request |
| Qwen3.6 27Bqwen/qwen3.6-27b | 27.8B | 1x A100 | About 15 min | $2.25/hour | $0.223 in, $2.30 out | Request |
| Granite 4.2 8Bibm-granite/granite-4.2-8b | 8.8B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Qwen3 14Bqwen/qwen3-14b | 14.8B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Ministral 3 14B 2512mistralai/ministral-14b-2512 | 13.9B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Nex-N2.5-Mininex-agi/nex-n2.5-mini | 35.1B (3.4B active) | 2x L40S | About 15 min | $2.75/hour | Quoted | Request |
| Qwen3 Coder Nextqwen/qwen3-coder-next | 79.7B (3.6B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.045 in, $0.303 out | Request |
| Qwen3 Coder 30B A3B Instructqwen/qwen3-coder-30b-a3b-instruct | 30.5B (3B active) | 1x L40S | About 15 min | $1.40/hour | $0.073 in, $0.294 out | Request |
| GLM 4.5 Airz-ai/glm-4.5-air | 111B (16.4B active) | 2x A100 | About 15 min | $4.50/hour | Quoted | Request |
| Qwen3.5-122B-A10Bqwen/qwen3.5-122b-a10b | 125B (12B active) | 1x RTX PRO 6000 | About 15 min | $3.15/hour | $0.157 in, $1.30 out | Request |
| Qwen3.5-27Bqwen/qwen3.5-27b | 27.8B | 1x A100 | About 15 min | $2.25/hour | $0.297 in, $2.40 out | Request |
| Phi 4microsoft/phi-4 | 14.7B | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Voxtral Small 24B 2507mistralai/voxtral-small-24b-2507 | 24.3B | 2x L40S | About 15 min | $2.75/hour | Quoted | Request |
| Qwen3 30B A3Bqwen/qwen3-30b-a3b | 30.5B (3B active) | 1x L40S | About 15 min | $1.40/hour | $0.169 in, $0.704 out | Request |
| Ling 3.0 Flash VLinclusionai/ling-3.0-flash-vl | 125B (5.5B active) | 2x A100 | About 15 min | $4.50/hour | Quoted | Request |
| Llama 3.2 1B Instructmeta-llama/llama-3.2-1b-instruct | 1.2B | 1x L40S | About 15 min | $1.40/hour | $0.01 in, $0.051 out | Request |
| Uncensoredcognitivecomputations/dolphin-mistral-24b-venice-edition | 24B | 2x L40S | About 15 min | $2.75/hour | Quoted | Request |
| Nemotron 3 Nano 30B A3Bnvidia/nemotron-3-nano-30b-a3b | 31.6B (2.6B active) | 1x L40S | About 15 min | $1.40/hour | Quoted | Request |
| Llama 3.2 3B Instructmeta-llama/llama-3.2-3b-instruct | 3.2B | 1x L40S | About 15 min | $1.40/hour | $0.016 in, $0.107 out | Request |
| Mistral Small 3.1 24Bmistralai/mistral-small-3.1-24b-instruct | 24B | 1x A100 | About 15 min | $2.25/hour | $0.463 in, $0.731 out | Request |
| Hy-MT2-1.8Btencent/hy-mt2-1.8b | 2B | 1x L40S | About 15 min | $1.40/hour | $0.011 in, $0.045 out | Request |
| UI-TARS 7Bbytedance/ui-tars-1.5-7b | 8.3B | 1x L40S | About 15 min | $1.40/hour | $0.041 in, $0.082 out | Request |
| Hy-MT2-7Btencent/hy-mt2-7b | 8B | 1x L40S | About 15 min | $1.40/hour | $0.033 in, $0.13 out | Request |
| LFM2.5-24B-A2Bliquid/lfm-2.5-24b-a2b | 24B (2.5B active) | 1x L40S | About 15 min | $1.40/hour | $0.072 in, $1.10 out | Request |
| Qwen2.5 Coder 32B Instructqwen/qwen-2.5-coder-32b-instruct | 32.8B | 1x A100 | About 15 min | $2.25/hour | Quoted | Request |
| Nemotron 3.5 Content Safetynvidia/nemotron-3.5-content-safety | 4.3B (3.6B active) | 1x L40S | About 15 min | $1.40/hour | $0.026 in, $0.026 out | Request |
| GLM 4.5Vz-ai/glm-4.5v | 108B (13.7B active) | 2x A100 | About 15 min | $4.50/hour | $0.593 in, $1.80 out | Request |
| Gemma 2 27Bgoogle/gemma-2-27b-it | 27.2B | 2x L40S | About 15 min | $2.75/hour | Quoted | Request |
| Hunyuan A13B Instructtencent/hunyuan-a13b-instruct | 80.4B (12.7B active) | 2x A100 | About 15 min | $4.50/hour | Quoted | Request |
| Qwen3 VL 30B A3B Thinkingqwen/qwen3-vl-30b-a3b-thinking | 31.1B (3.6B active) | 1x L40S | About 15 min | $1.40/hour | $0.27 in, $0.929 out | Request |
| LFM2.5-2.6Bliquid/lfm-2.5-2.6b | 2.7B | 1x L40S | About 15 min | $1.40/hour | $0.057 in, $0.082 out | Request |
| Reka Flash 3rekaai/reka-flash-3 | 20.9B | 2x L40S | About 15 min | $2.75/hour | Quoted | Request |
| Qwen2.5 7B Instructqwen/qwen-2.5-7b-instruct | 7.6B | 1x L40S | About 15 min | $1.40/hour | $0.041 in, $0.082 out | Request |
$5 to $20 an hour
Larger models across two to four GPUs.
| Model | Size | Runs on | Ready in | Reserved | Per million tokens | Request |
|---|---|---|---|---|---|---|
| DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731 | 304B (13B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| DeepSeek V4 Flash 0423deepseek/deepseek-v4-flash | 291B (13B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| MiniMax M3.1 Flashminimax/minimax-m3.1-flash | 221B (10B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| MiMo-V2.6-Flashxiaomi/mimo-v2.6-flash | 311B (16.8B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| Hy3tencent/hy3 | 299B (23.9B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| Ling 3.0 Flashinclusionai/ling-3.0-flash | 128B (8.2B active) | 4x L40S | About 15 min | $5.45/hour | Quoted | Request |
| Llama 3.3 70B Instructmeta-llama/llama-3.3-70b-instruct | 70.6B | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| Qwen3 235B A22B Instruct 2507qwen/qwen3-235b-a22b-2507 | 235B (21.6B active) | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| MiniMax M3minimax/minimax-m3 | 427B (25.6B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| Step 3.7 Flashstepfun/step-3.7-flash | 201B (15.9B active) | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| Qwen3 VL 235B A22B Instructqwen/qwen3-vl-235b-a22b-instruct | 236B (22.1B active) | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| Nemotron 3 Supernvidia/nemotron-3-super-120b-a12b | 124B (11.5B active) | 4x L40S | About 15 min | $5.45/hour | Quoted | Request |
| Qwen2.5 72B Instructqwen/qwen-2.5-72b-instruct | 72.7B | 1x H200 | About 15 min | $6.65/hour | Quoted | Request |
| Qwen3.5 397B A17Bqwen/qwen3.5-397b-a17b | 403B (23.4B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | $0.385 in, $2.40 out | Request |
| DeepSeek V4 Flash Vision Expdeepseek/deepseek-v4-flash-vision-exp | 305B (13B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| Qwen2.5 VL 72B Instructqwen/qwen2.5-vl-72b-instruct | 73.4B | 1x H200 | About 15 min | $6.65/hour | Quoted | Request |
| Llama 3.1 70B Instructmeta-llama/llama-3.1-70b-instruct | 70.6B | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| Hy3 previewtencent/hy3-preview | 299B (23.9B active) | 4x RTX PRO 6000 | About 30 min | $12.45/hour | Quoted | Request |
| Qwen3 Coder 480B A35Bqwen/qwen3-coder | 480B (34.5B active) | 4x RTX PRO 6000 | About 20 min | $12.45/hour | Quoted | Request |
| Hermes 3 70B Instructnousresearch/hermes-3-llama-3.1-70b | 70.6B | 1x H200 | About 15 min | $6.65/hour | Quoted | Request |
| Devstral 2 2512mistralai/devstral-2512 | 125B | 1x B200 | About 20 min | $10.00/hour | Quoted | Request |
| GLM 4.6Vz-ai/glm-4.6v | 108B (13.7B active) | 4x L40S | About 20 min | $5.45/hour | Quoted | Request |
| Step 3.5 Flashstepfun/step-3.5-flash | 199B (13.9B active) | 4x A100 | About 20 min | $8.95/hour | Quoted | Request |
| Magnum v4 72Banthracite-org/magnum-v4-72b | 72.7B | 1x H200 | About 15 min | $6.65/hour | $1.80 in, $3.60 out | Request |
| Qwen3 235B A22B Thinking 2507qwen/qwen3-235b-a22b-thinking-2507 | 235B (21.6B active) | 2x RTX PRO 6000 | About 15 min | $6.25/hour | Quoted | Request |
| MiniMax M2.1minimax/minimax-m2.1 | 229B (10.4B active) | 4x A100 | About 20 min | $8.95/hour | Quoted | Request |
| MiniMax M2minimax/minimax-m2 | 229B (10.4B active) | 4x A100 | About 20 min | $8.95/hour | Quoted | Request |
| Mixtral 8x22B Instructmistralai/mixtral-8x22b-instruct | 141B (39B active) | 4x L40S | About 15 min | $5.45/hour | $3.00 in, $9.00 out | Request |
Over $20 an hour
The largest open models, on four to eight data-center GPUs.
| Model | Size | Runs on | Ready in | Reserved | Per million tokens | Request |
|---|---|---|---|---|---|---|
| DeepSeek V4.1 Flash 1017deepseek/deepseek-v4.1-flash-1017 | 763B | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| GLM 5.3 Flashz-ai/glm-5.3-flash | 321B (24.7B active) | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flash | 763B | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Hy4 previewtencent/hy4-preview | 780B (49B active) | 4x B300 | 65 min or more | $44.95/hour | $1.10 in, $3.20 out | Request |
| GLM 5.3z-ai/glm-5.3 | 753B (50.3B active) | 4x B300 | 40 min or more | $44.95/hour | $0.741 in, $2.40 out | Request |
| DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813 | 1651B (49B active) | 4x B300 | 40 min or more | $44.95/hour | Quoted | Request |
| DeepSeek V3.2deepseek/deepseek-v3.2 | 685B (51B active) | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Kimi K3moonshotai/kimi-k3 | 2780B (105B active) | 8x B300 | 65 min or more | $89.90/hour | $2.40 in, $12.00 out | Request |
| GLM 5.2z-ai/glm-5.2 | 753B (50.3B active) | 4x B300 | 65 min or more | $44.95/hour | $0.804 in, $2.60 out | Request |
| Step 5 Previewstepfun/step-5-preview | 604B (32.4B active) | 4x B200 | About 25 min | $39.95/hour | $1.40 in, $3.80 out | Request |
| DeepSeek V4 Pro 0423deepseek/deepseek-v4-pro | 1599B (49B active) | 4x B300 | 40 min or more | $44.95/hour | Quoted | Request |
| MiMo-V2.6-Proxiaomi/mimo-v2.6-pro | 1024B (43.9B active) | 4x B200 | About 30 min | $39.95/hour | Quoted | Request |
| DeepSeek V3.1deepseek/deepseek-chat-v3.1 | 685B (50.1B active) | 4x B200 | About 25 min | $39.95/hour | $0.329 in, $0.986 out | Request |
| DeepSeek V3deepseek/deepseek-chat | 685B (50.1B active) | 4x B300 | 35 min or more | $44.95/hour | Quoted | Request |
| Llama 4 Maverickmeta-llama/llama-4-maverick | 402B | 2x B200 | About 20 min | $20.00/hour | Quoted | Request |
| Kimi K2.6moonshotai/kimi-k2.6 | 1027B (32.2B active) | 4x B200 | About 30 min | $39.95/hour | $0.99 in, $4.30 out | Request |
| MiMo-V2.5xiaomi/mimo-v2.5 | 311B (16.8B active) | 4x B200 | About 25 min | $39.95/hour | $0.338 in, $1.20 out | Request |
| DeepSeek V3 0324deepseek/deepseek-chat-v3-0324 | 685B (50.1B active) | 4x B200 | About 25 min | $39.95/hour | $0.397 in, $1.60 out | Request |
| Nemotron 3 Ultranvidia/nemotron-3-ultra-550b-a55b | 561B (53.9B active) | 4x B200 | 50 min or more | $39.95/hour | $0.608 in, $2.70 out | Request |
| GLM 4.7z-ai/glm-4.7 | 358B (38.4B active) | 4x B200 | About 35 min | $39.95/hour | Quoted | Request |
| GLM 5z-ai/glm-5 | 754B (50.8B active) | 4x B300 | 65 min or more | $44.95/hour | $0.982 in, $3.20 out | Request |
| Kimi K2.5moonshotai/kimi-k2.5 | 1027B (32.2B active) | 4x B200 | About 30 min | $39.95/hour | $0.721 in, $3.70 out | Request |
| GLM 5.1z-ai/glm-5.1 | 754B (50.8B active) | 4x B300 | 65 min or more | $44.95/hour | $1.10 in, $3.30 out | Request |
| DeepSeek V3.2 Expdeepseek/deepseek-v3.2-exp | 685B (51B active) | 4x B300 | 35 min or more | $44.95/hour | Quoted | Request |
| Qwen3.8 2.4T A95Bqwen/qwen3.8-2.4t-a95b | 2446B (120B active) | 8x B300 | 60 min or more | $89.90/hour | $2.10 in, $6.30 out | Request |
| Kimi K2 0905moonshotai/kimi-k2-0905 | 1026B (31.7B active) | 8x B200 | 45 min or more | $79.90/hour | $0.84 in, $3.50 out | Request |
| Kimi K2.7 Codemoonshotai/kimi-k2.7-code | 1027B (32.2B active) | 4x B200 | About 30 min | $39.95/hour | $1.30 in, $5.20 out | Request |
| R1 0528deepseek/deepseek-r1-0528 | 685B (50.1B active) | 4x B200 | About 25 min | $39.95/hour | $0.418 in, $1.80 out | Request |
| MiMo-V2.5-Proxiaomi/mimo-v2.5-pro | 1023B (42.9B active) | 8x B200 | 45 min or more | $79.90/hour | Quoted | Request |
| DeepSeek V3.1 Terminusdeepseek/deepseek-v3.1-terminus | 685B (50.1B active) | 4x B300 | 35 min or more | $44.95/hour | Quoted | Request |
| Inkling Smallthinkingmachines/inkling-small | 266B | 2x B200 | About 20 min | $20.00/hour | Quoted | Request |
| GLM 4.6z-ai/glm-4.6 | 357B (36.8B active) | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Kimi K2 0711moonshotai/kimi-k2 | 1026B (31.7B active) | 8x B200 | 45 min or more | $79.90/hour | Quoted | Request |
| Nex-N2.5-Pronex-agi/nex-n2.5-pro | 397B (16.8B active) | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Hermes 4 405Bnousresearch/hermes-4-405b | 406B | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Inklingthinkingmachines/inkling | 952B | 8x B200 | 30 min or more | $79.90/hour | Quoted | Request |
| GLM 4.5z-ai/glm-4.5 | 358B (38.4B active) | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Hermes 3 405B Instructnousresearch/hermes-3-llama-3.1-405b | 406B | 4x B200 | About 25 min | $39.95/hour | Quoted | Request |
| Kimi K2 Thinkingmoonshotai/kimi-k2-thinking | 1026B (31.7B active) | 4x B200 | About 30 min | $39.95/hour | Quoted | Request |
Not on the list?
Send us the Hugging Face link. If it has open weights, generates text and its license allows serving it as an API, we can usually run it. Prices on this page are from November 1, 2026 and follow what the GPUs cost us.