Same open weights, served by 3 providers. Weights: Qwen/Qwen2.5-7B-Instruct.
| Provider | Their model id | Serving | Input | Output | Cached in | Blended 3:1 | Source | |
|---|---|---|---|---|---|---|---|---|
DeepInfra Deprecated, 2025-11-06 | Qwen/Qwen2.5-7B-Instruct | standardFull precision | $0.04 | $0.10 | โ | $0.055 | Verified 12h ago ยท Changed 40d ago | |
Together AI Deprecated | Qwen/Qwen2.5-7B-Instruct-Turbo | standardQuantization not disclosed | $0.30 | $0.30 | โ | $0.30 | Verified 12h ago ยท Changed 40d ago | |
Alibaba Cloud Model Studio Original creatorDeprecated | qwen2-5-7b-instruct | standardQuantization not disclosed | $0.175 | $0.70 | โ | $0.306 | Verified 31d ago ยท Changed 40d ago Not verified in 736h |
Same weights don't always mean identical outputs: quantization, context limits and serving stacks can change quality, especially for tool calling. Prices are USD per 1M tokens; hyperscaler prices are for the reference region shown.
| DeepInfra bfloat16 | $6/mo |
| Alibaba Cloud Model Studio | $31.5/mo |
| Together AI | $36/mo |
Cache hits use each provider's cached-input price where it has one. Cache write costs aren't included.