Model library
Every entry is built from the published repository: parameter count, architecture, context length, licence, and the memory the weights and the cache actually need. 82 models listed.
| Model | Size | Context | Licence | Weights | Single card |
|---|---|---|---|---|---|
| tencent/Hy-Memory-30B-A3B | 30.1B | 262,144 | apache-2.0 | 56 GiB | A100 80GB |
| ibm-granite/granite-switch-4.1-8b | 8.8B | 131,072 | apache-2.0 | 16 GiB | RTX 4090 24GB |
| ibm-granite/granite-switch-4.1-30b | 30B | 131,072 | apache-2.0 | 56 GiB | A100 80GB |
| sabicap/sabi-noema-mla | 27.4B | 262,144 | apache-2.0 | 51 GiB | A100 80GB |
| inclusionAI/Ling-3.1-flash | 572.1B | 262,144 | mit | 534 GiB | multi-GPU |
| nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-LiLiCorr | 673M | 262,144 | other | 1 GiB | RTX 4090 24GB |
| LiquidAI/LFM2.5-24B-A2B | 24B | 131,072 | other | 45 GiB | A100 80GB |
| LiquidAI/LFM2.5-24B-A2B-Base | 24B | 131,072 | other | 45 GiB | A100 80GB |
| reflection/Beam-1.0-501B-A23B | 501.3B | 1,048,576 | apache-2.0 | 934 GiB | multi-GPU |
| rockmatrix-ai/EdgeRock | 4.6B | 262,144 | other | 9 GiB | RTX 4090 24GB |
| stonehill-tech/EdgeRock | 4.6B | 262,144 | other | 9 GiB | RTX 4090 24GB |
| tventurella/Mr.Chatterbox-2.7B-Chat | 2.7B | 2,048 | cc-by-nc-sa-4.0 | 10 GiB | RTX 4090 24GB |
| Blockway/Agens-Volundr-32B-Preview | 32.7B | 262,144 | apache-2.0 | 61 GiB | A100 80GB |
| amazon/ALoDLM-8B | 8.2B | 40,960 | cc-by-nc-4.0 | 15 GiB | RTX 4090 24GB |
| Aleph-Alpha/Kolibri-1 | 78.1B | 262,144 | apache-2.0 | 73 GiB | RTX PRO 6000 96GB |
| InternScience/MLExpert | 34.7B | 262,144 | apache-2.0 | 65 GiB | A100 80GB |
| IQuestLab/IQuest-Q1 | 320.3B | 524,288 | other | 597 GiB | multi-GPU |
| XiaomiMiMo/MiMo-V2.6-Flash-MOPD | 310.8B | 1,048,576 | mit | 161 GiB | multi-GPU |
| XiaomiMiMo/MiMo-V2.6-Pro-MOPD | 1024.2B | 1,048,576 | mit | 527 GiB | multi-GPU |
| FINAL-Bench/Darwin-27B-RSI | 26.9B | 262,144 | apache-2.0 | 50 GiB | A100 80GB |
| IFM/K2-Type-0.9B | 1.1B | 131,072 | apache-2.0 | 2 GiB | RTX 4090 24GB |
| NovelAI/clio-v1-legacy | 3B | 8,192 | gpl-2.0 | 6 GiB | RTX 4090 24GB |
| cosmicoptima/computer-10 | 70.6B | 131,072 | llama3.1 | 131 GiB | multi-GPU |
| paradigma-inc/limite-1b-violetto | 1B | 131,072 | apache-2.0 | 2 GiB | RTX 4090 24GB |
| XiaomiMiMo/MiMo-V2.6-Flash-RL | 159.4B | 1,048,576 | mit | 86 GiB | H200 141GB |
| XiaomiMiMo/MiMo-V2.6-Pro-RL | 524.1B | 1,048,576 | mit | 280 GiB | multi-GPU |
| JetBrains/Mellum2.1-12B-A2.5B-Thinking | 12.1B | 131,072 | apache-2.0 | 23 GiB | RTX 5090 32GB |
| XingChen-AGI/Xing4.0-29B-A4B | 31.2B | 262,144 | apache-2.0 | 58 GiB | A100 80GB |
| llm-jp/llm-jp-4.1-33b-thinking | 33.2B | 65,536 | apache-2.0 | 62 GiB | A100 80GB |
| llm-jp/llm-jp-4.1-32b-a3b-thinking | 32.1B | 65,536 | apache-2.0 | 60 GiB | A100 80GB |
| Gensyn/open-1b-sft | 1.6B | 4,096 | apache-2.0 | 6 GiB | RTX 4090 24GB |
| yandex/AliceAI-Foundation-80B-A3B-Base | 81.3B | 262,144 | apache-2.0 | 151 GiB | multi-GPU |
| nex-agi/Nex-N2.5-Pro | 396.8B | 262,144 | apache-2.0 | 379 GiB | multi-GPU |
| Gryphe/Pantheon-Reasoning-26B-A4B-1.1-V2 | 26.5B | 262,144 | apache-2.0 | 49 GiB | A100 80GB |
| nex-agi/Nex-N2.5-mini | 35.1B | 262,144 | apache-2.0 | 65 GiB | A100 80GB |
| zgcagi/ZGCM-1-7B | 7.4B | 262,144 | mit | 14 GiB | RTX 4090 24GB |
| nex-agi/Nex-N2.5-Max | 1600.8B | 1,048,576 | apache-2.0 | 1541 GiB | multi-GPU |
| openbmb/MiniCPM5-2B | 2.5B | 131,072 | apache-2.0 | 5 GiB | RTX 4090 24GB |
| TokenRhythm/NeoHorse-1-9B | 9B | 262,144 | apache-2.0 | 17 GiB | RTX 4090 24GB |
| TokenRhythm/NeoHorse-1-4B | 4.2B | 262,144 | apache-2.0 | 8 GiB | RTX 4090 24GB |
| inclusionAI/LLaDA2.2-mini | 16.3B | 131,072 | apache-2.0 | 30 GiB | L40S 48GB |
| ai-sage/GigaChat3.5-432B-A28B-Reasoning | 438.1B | 262,144 | mit | 415 GiB | multi-GPU |
| inclusionAI/Ling-3.0-flash-Fin | 127.5B | 262,144 | mit | 237 GiB | multi-GPU |
| IFM/K2-Horizon-375B-A23B | 379.2B | 524,288 | apache-2.0 | 706 GiB | multi-GPU |
| IFM/K2-Horizon-7B | 9B | 524,288 | apache-2.0 | 17 GiB | RTX 4090 24GB |
| IFM/K2-Horizon-3.7B | 5.1B | 524,288 | apache-2.0 | 9 GiB | RTX 4090 24GB |
| IFM/K2-Horizon-32B | 34.8B | 524,288 | apache-2.0 | 65 GiB | A100 80GB |
| IFM/K2-Horizon-MoVA-36B-A4B | 37.4B | 524,288 | apache-2.0 | 70 GiB | RTX PRO 6000 96GB |
| IFM/K2-Horizon-0.9B | 1.1B | 131,072 | apache-2.0 | 2 GiB | RTX 4090 24GB |
| openbmb/MiniCPM5-2B-Midtrain | 2.5B | 131,072 | apache-2.0 | 5 GiB | RTX 4090 24GB |
| Nanbeige/Nanbeige4.2-3B-DSpark | 848M | 262,144 | apache-2.0 | 2 GiB | RTX 4090 24GB |
| menezesbruno/manaca-1b-base | 1.7B | 4,096 | cc-by-4.0 | 3 GiB | RTX 4090 24GB |
| XHToken/Spark-X2.5-4B-Base | 4.1B | 1,048,576 | apache-2.0 | 8 GiB | RTX 4090 24GB |
| XHToken/Spark-X2.5-1.7B-Base | 1.7B | 1,048,576 | apache-2.0 | 3 GiB | RTX 4090 24GB |
| openbmb/MiniCPM5-2B-SFT | 2.5B | 131,072 | apache-2.0 | 5 GiB | RTX 4090 24GB |
| openbmb/MiniCPM5-2B-Base | 2.5B | 524,288 | apache-2.0 | 5 GiB | RTX 4090 24GB |
| tencent/Hy4-preview | 780B | 1,048,576 | apache-2.0 | 1453 GiB | multi-GPU |
| zai-org/GLM-5.3 | 753.3B | 1,048,576 | other | 704 GiB | multi-GPU |
| pipecat-ai/phonellm-alpha-1 | 31.6B | 262,144 | bsd-2-clause | 59 GiB | A100 80GB |
| XHToken/Spark-X2.5-4B | 4.1B | 1,048,576 | apache-2.0 | 8 GiB | RTX 4090 24GB |
Hosting
Send us the model name and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint. A model does not have to be on this list.