LLM Reference

Qwen2-72B

Released
2024-06-05
Last refreshed
2026-05-16
Status
Researched 46d ago
CodingRAGLong contextClassificationJSON / Tool use

Qwen2-72B is worth evaluating for coding, rag, and long context when its provider route and context window match the workload.

Use it for

  • Teams evaluating coding, rag, and long context
  • Workloads that can use a 128k context window
  • Buyers comparing 4 tracked provider routes

Do not use it for

  • Vision or document-understanding workloads
Specifications
Family
Qwen2
Released
2024-06-05
Context
128k
Parameters
72.71B
Architecture
Decoder Only
Specialization
general
Training
finetuned
Created by

AI research institute of Alibaba Group.

Hangzhou, Zhejiang, China
Founded 2017
Website
Pricing
Output / 1M
$0.650
Input / 1M
$0.450

Cheapest of 4 routes · DeepInfra

About

Qwen2-72B is a cutting-edge large language model developed by Alibaba's Qwen team, featuring an impressive 72 billion parameters based on the Transformer architecture 12. It employs advanced enhancements such as SwiGLU activation, attention QKV bias, and group query attention to advance efficiency and precision 16. The model demonstrates strong performance across diverse benchmarks, excelling in language understanding, generation, coding, mathematics, and multilingual tasks, often surpassing other open-source models and challenging proprietary alternatives 34. With support for processing up to 128,000 tokens in context and proficiency in around 30 languages, it offers extensive input capabilities 15. However, the base model is not optimal for direct text generation; post-training techniques are advisable for specific applications 16.

Qwen2-72B is a model in the Qwen2 family. The structured metadata tracks a 128k-token context window and structured outputs. This page tracks provider routes through Fireworks AI, DeepInfra, Together AI, and 1 more, with the cheapest tracked route listed at $0.45 input and $0.65 output per 1M tokens. Headline tracked benchmarks include HumanEval 67.1, Massive Multitask Language Understanding 84.2, and MMLU PRO 64.4.

Top use-case fit: coding, agents, and build tasks

Coding

Q/$ C

1 relevant benchmark in the decision map.

RAG

Included by capability and metadata signals in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Provider price ladder

Compare all 4

Compare API pricing across 4 providers for input and output tokens, batch, and cached reads when available.

ProviderInput / 1MOutput / 1MRoute
DeepInfra$0.450$0.650
Serverless
Fireworks AI$0.900$0.900
Serverless
Together AI$0.900$0.900
Serverless
Microsoft Foundry$1.00$2.00
Provisioned

Capabilities

Structured Outputs

Benchmark peer barsfor Coding

Benchmark scores(3)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionSource
HumanEval67.1pass@1https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard
Massive Multitask Language Understanding84.25-shothttps://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard
MMLU PRO64.4https://huggingface.co/spaces/TIGER-Lab/MMLU-Pro

Migration checks

No linked migration route is available for this model yet.

Rankings & picks(10)

Comparison and alternatives

Browse all comparisons →