Qwen2.5 72B Instruct
Qwen2.5 72B InstructAlternatives?
CheaperMiMo v2.6 ProthinkingOpen×13.9-78%$0.13Memory (4-bit): 44 GB → 615 GB+4
BetterGPT-6 Lunalow×5.7-92%$0.005Function calling only with none · Chat Completions+3OpenQwen3.8 Flashthinking×10.7-91%$0.37Memory (4-bit): 44 GB → 108 GB+6
CheaperMuse Spark 1.3max×13.7-18%$1.60Only auto tool_choice is supported+5
CheaperMiniMax M3.1 Flashthinking×13.5-89%$0.22Complete model response must be appendedMiMo v2.6 ProthinkingOpen×13.9-78%$0.13Memory (4-bit): 44 GB → 615 GB+4Muse Spark 1.3max×13.7-18%$1.60Only auto tool_choice is supported+5MiniMax M3.1 Flashthinking×13.5-89%$0.22Complete model response must be appended
Gemini 3.8 Flashhigh×13.4-39%*$1.24Sampling parameters are ignored+1
Step 5 PreviewthinkingOpen×13.1-42%$1.03Memory (4-bit): 44 GB → 363 GB+3GPT-6 Lunalow×5.7-92%$0.005Function calling only with none · Chat Completions+3
Granite 4.2 3BthinkingOpen+86%-98%$0.006Context: 131K → 128KGPT-5.6 Lunalow×5.6-82%$0.01GPT OSS 20BhighOpen×3.1-99%$0.011
Mistral Small 4thinkingOpen×2.8-89%$0.015Memory (4-bit): 44 GB → 72 GBQwen3.8 FlashthinkingOpen×10.7-91%$0.37Memory (4-bit): 44 GB → 108 GB+6Qwen3.8 27BxhighOpen×9.5-61%$1.01max_tokens range · QwenCloud+4
Nemotron 3.5 Super VLOpen×8.6-93%$0.19Memory (4-bit): 44 GB → 75 GBQwen 3.5 122B A10BthinkingOpen×7.1-55%$0.32Memory (4-bit): 44 GB → 75 GBNemotron 3.5 Bolt 74B-A9BOpen×6.9——Memory (4-bit): 44 GB → 45 GBQwen3.8 Max (0902)thinking×12.1—$5.41Qwen3.8 2.4T A95BthinkingOpen×11.9—$2.16Memory (4-bit): 44 GB → 1,468 GB+5Qwen3.7 Maxthinking×11.4—$1.15Qwen3.8 FlashthinkingOpen×10.7—$0.37Memory (4-bit): 44 GB → 108 GB+6Qwen3.7 Plusthinking×10—$0.22Qwen3.6 Max Previewthinking×8.7—$1.30 / $7.80Include the word JSON · QwenCloudQwen3.5 397B A17BthinkingOpen×8.1—$0.47Memory (4-bit): 44 GB → 242 GBQwen3 Max Thinkingthinking×7.9—$1.20 / $6All Alibaba (Qwen) models →Run it yourself
estimateBenchmarks
in Chat & KnowledgeFAQ
How much does Qwen2.5 72B Instruct cost?›
$1.40 per 1M input tokens and $5.60 per 1M output tokens; cheapest via DeepInfra at $0.36 / $0.40.
Is Qwen2.5 72B Instruct free to use?›
There is no free API; the cheapest paid price is $0.36 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Qwen2.5 72B Instruct, and what is it best at?›
#323 of 334 in Chat & Knowledge, #161 of 239 in Coding, #135 of 242 in Agents & Tool Use, #158 of 195 in Writing & Language.
Is Qwen2.5 72B Instruct better than Qwen3 VL 4B Instruct?›
On Humanity's Last Exam, Qwen2.5 72B Instruct scores 3.6 vs 3.6 — about the same.
How can I use Qwen2.5 72B Instruct?›
Through the API of DeepInfra.
What is the context window of Qwen2.5 72B Instruct?›
131K tokens, with up to 8K output tokens.
When was Qwen2.5 72B Instruct released? Is there a newer version?›
Released 2024-09-16. It has been superseded by Qwen3 235B A22B.
Is Qwen2.5 72B Instruct open source?›
Yes, the weights are open (qwen licence).
Can I run Qwen2.5 72B Instruct locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Qwen2.5 72B Instruct read images?›
No, it accepts text only.
What is Qwen2.5 72B Instruct?›
Qwen2.5 72B Instruct is an open-weights language model from Alibaba (Qwen), released Sep 2024.
What are the best alternatives to Qwen2.5 72B Instruct?›
MiMo v2.6 Pro (cheaper), GPT-6 Luna (better), Qwen3.8 Flash (open weights).