Open source AI models
375 open-weight models you can download or call through any host, ranked by quality.
/
 Phi 4 MultimodalMicrosoftMITVision | $0.48/1Mvia AIHubMix | 128K | $0.12/1M | $0.48/1M | — | Feb 24, 2025 | |
| Phi 4 Reasoning PlusMicrosoftMIT | $0.50/1Mvia Azure | — | $0.125/1M | $0.50/1M | — | Apr 17, 2025 | |
| Phi-3-medium-4k-instructMicrosoftMIT | $0.68/1Mvia Azure | — | $0.17/1M | $0.68/1M | — | May 7, 2024 | |
 Qianfan OcrBaiduApache 2.0Vision | $0.248/1Mvia AIHubMix | 32K | $0.062/1M | $0.248/1M | — | Mar 18 | |
 Qwen 2.5 VL 32B InstructAlibabaApache 2.0JSON+2 | $0.24/1Mvia AIHubMix | — | $0.24/1M | $0.24/1M | — | Mar 21, 2025 | |
 Qwen-SEA-LION-v4-32B-ITAI SingaporeMIT | $0.50/1Mvia Hugging Face | — | $0.25/1M | $0.50/1M | — | Oct 16, 2025 | |
| Qwen-SEA-LION-v4.5-27B-ITAI SingaporeMIT | $0.40/1Mvia Hugging Face | — | $0.20/1M | $0.40/1M | — | May 14 | |
| Qwen2-72B-InstructAlibabaOpen weights · with terms | $0.80/1Mvia AIHubMix | — | $0.80/1M | $0.80/1M | — | May 22, 2024 | |
| Qwen2-VL-72B-InstructAlibabaOpen weights · with termsVision | $6.54/1Mvia AIHubMix | — | $2.18/1M | $6.54/1M | — | Sep 17, 2024 | |
| Qwen2-VL-7B-InstructAlibabaApache 2.0Vision | $0.70/1Mvia AIHubMix | — | $0.28/1M | $0.70/1M | — | Aug 28, 2024 | |
| Qwen2.5 Math 1.5b InstructAlibabaApache 2.0 | $0.20/1Mvia AIHubMix | — | $0.20/1M | $0.20/1M | — | Sep 16, 2024 | |
| Qwen2.5 Math 7B InstructAlibabaApache 2.0 | $0.40/1Mvia AIHubMix | — | $0.20/1M | $0.40/1M | — | Sep 19, 2024 | |
| Qwen2.5-32B-InstructAlibabaApache 2.0 | $0.175/1Mvia SiliconFlow | — | $0.175/1M | $0.175/1M | — | Sep 15, 2024 | |
| Qwen2.5-72B-InstructAlibabaOpen weights · with termsJSON+1 | $0.40/1Mvia Deep Infra | 33K | $0.36/1M | $0.40/1M | — | Sep 16, 2024 | |
| Qwen2.5-Coder-32B-InstructAlibabaApache 2.0JSON | $0.16/1Mvia AIHubMix | 33K | $0.16/1M | $0.16/1M | — | Nov 6, 2024 | |
| Qwen2.5-VL 72B InstructAlibabaOpen weights · with termsPrompt caching+2 | $0.50/1Mvia AIHubMix | 128K | $0.50/1M | $0.50/1M | — | Jan 27, 2025 | |
| Qwen2p5 14BAlibabaApache 2.0 | $0.097/1Mvia SiliconFlow | — | $0.097/1M | $0.097/1M | — | Sep 15, 2024 | |
| Qwen2p5 Coder 1.5BAlibabaApache 2.0 | $0.40/1Mvia AIHubMix | — | $0.20/1M | $0.40/1M | — | Sep 18, 2024 | |
| Qwen2p5 Coder 3BAlibabaNon-commercial | $0.03/1Mvia Hugging Face | — | $0.01/1M | $0.03/1M | — | Nov 8, 2024 | |
| Qwen2p5 Math 72B InstructAlibabaOpen weights · with terms | $2.40/1Mvia AIHubMix | — | $0.80/1M | $2.40/1M | — | Sep 16, 2024 | |
| Qwen3 235B A22BAlibabaApache 2.0JSON+2 | $0.60/1Mvia Hugging Face | 128K | $0.20/1M | $0.60/1M | — | Apr 27, 2025 | |
| Qwen3 235B A22B Instruct 2507AlibabaApache 2.0JSON+1 | $0.10/1Mvia Requesty | 262K | $0.071/1M | $0.10/1M | — | Jul 21, 2025 | |
| Qwen3 235B A22b Thinking 2507AlibabaApache 2.0JSON+2 | $2.30/1Mvia Hugging Face | 262K | $0.23/1M | $2.30/1M | — | Jul 25, 2025 | |
| Qwen3 30B A3B Thinking 2507AlibabaApache 2.0 | $1.20/1Mvia AIHubMix | — | $0.12/1M | $1.20/1M | — | Jul 29, 2025 | |
| Qwen3 30B-A3B Instruct 2507AlibabaApache 2.0JSON+1 | $0.30/1Mvia Cortecs | 262K | $0.10/1M | $0.30/1M | — | Jul 28, 2025 | |
| Qwen3 Next 80B A3B InstructAlibabaApache 2.0JSON+1 | $1.10/1Mvia Deep Infra | 262K | $0.09/1M | $1.10/1M | — | Sep 9, 2025 | |
| Qwen3 Next 80B A3B ThinkingAlibabaApache 2.0JSON+2 | $1.20/1Mvia OpenRouter | 262K | $0.15/1M | $1.20/1M | — | Sep 9, 2025 | |
| Qwen3 Omni 30B A3B ThinkingAlibabaOpen weights · permissive | $0.389/1Mvia SiliconFlow | — | $0.097/1M | $0.389/1M | — | Sep 15, 2025 | |
| Qwen3 VL 235B A22B InstructAlibabaApache 2.0Prompt caching+3 | $0.88/1Mvia Deep Infra | 262K | $0.20/1M | $0.88/1M | — | Sep 22, 2025 | |
| Qwen3 VL 235B A22B ThinkingAlibabaApache 2.0JSON+3 | $3.95/1Mvia DevPass (LLM Gateway) | 131K | $0.98/1M | $3.95/1M | — | Sep 22, 2025 | |
| Qwen3 VL 30B A3B InstructAlibabaApache 2.0JSON+3 | $0.389/1Mvia SiliconFlow | 262K | $0.097/1M | $0.389/1M | — | Sep 30, 2025 | |
| Qwen3-VL-32B-InstructAlibabaApache 2.0 | $0.556/1Mvia SiliconFlow | — | $0.139/1M | $0.556/1M | — | Oct 19, 2025 | |
| Qwen3-VL-32B-ThinkingAlibabaApache 2.0 | $1.39/1Mvia SiliconFlow | — | $0.139/1M | $1.39/1M | — | Oct 19, 2025 | |
| Qwen3.5 122B-A10BAlibabaApache 2.0JSON+3 | $0.901/1Mvia AIHubMix | 262K | $0.40/1M | $3.20/1M | — | Feb 24 | |
| Qwq 32BAlibabaApache 2.0JSON+1 | $0.56/1Mvia AIHubMix | — | $0.14/1M | $0.56/1M | — | Mar 5, 2025 | |
| QwQ-32B-PreviewAlibabaApache 2.0 | $0.16/1Mvia AIHubMix | — | $0.16/1M | $0.16/1M | — | Nov 27, 2024 | |
 Recurrentgemma 2BGoogleGemma | Free/1Mvia Nvidia | — | Free/1M | Free/1M | — | Apr 6, 2024 | |
 Ring 1tInclusionaiMITJSON+2 | $2.19/1Mvia AIHubMix | — | $0.548/1M | $2.19/1M | — | Oct 10, 2025 | |
| Ring Flash 2.0InclusionaiMITJSON+2 | $0.544/1Mvia AIHubMix | — | $0.136/1M | $0.544/1M | — | Sep 19, 2025 | |
| Ring-2.6-1TInclusionaiMITPrompt caching+3 | $1.10/1Mvia ZenMux | 262K | $0.132/1M | $1.10/1M | — | May 14 | |
 Riva Translate 4B InstructNVIDIAOpen weights · with terms | Free/1Mvia Nvidia | — | Free/1M | Free/1M | — | Jun 9, 2025 | |
| Qwen ImageAlibabaApache 2.0 | $0.02/imgvia FastRouter | 4K | — | $0.02/img | — | Aug 2, 2025 | |
| Qwen3 Embedding 8BAlibabaApache 2.0 | $0.039/1Mvia SiliconFlow | 33K | — | $0.039/1M | — | Jun 3, 2025 | |
| Qwen3-VL Embedding 8BAlibabaApache 2.0Vision | $0.097/1Mvia SiliconFlow | 8K | — | $0.097/1M | — | Jan 7 | |
| Qwen3 Reranker 8BAlibabaApache 2.0Vision | $0.039/1Mvia SiliconFlow | — | — | $0.039/1M | — | May 29, 2025 | |
| Qwen3-Reranker-4BAlibabaApache 2.0Vision | $0.019/1Mvia SiliconFlow | — | — | $0.019/1M | — | Jun 3, 2025 | |
| Qwen3-VL-Reranker-8BAlibabaApache 2.0 | $0.097/1Mvia SiliconFlow | — | — | $0.097/1M | — | Jan 7 | |
 SAM 3MetaOpen weights | $0.005/runvia Fal | — | — | $0.005/run | — | Nov 7, 2025 | |
| Privacy FilterOpenAIApache 2.0 | Free/1Mvia NEAR AI Cloud | 1K | $0.01/1M | Free/1M | — | Apr 17 | |
| Prompt Guard 2 86MMetaOpen weights · with terms | $0.01/1Mvia Helicone | — | $0.01/1M | $0.01/1M | — | Apr 28, 2025 |
Frequently asked questions
Ranked here by quality. Open models now trail the best closed ones by little, and the hosts that serve them often charge much less.
Most "open source" models publish their weights under a license, but not their training data. We list every model whose weights you can download.
On your own hardware, or through inference hosts that charge per token. Open a model's providers to compare each host's price, speed and uptime.
Can't find what you're looking for? Contact us