Qwen2.5 72B
Qwen2.5's 72B flagship: 128K context, 29 languages, leaderboard-topping.
About
Qwen2.5-72B-Instruct was Alibaba's open-weights flagship of late 2024: a 72B dense model with 128K-token context, 29 languages, strong coding/math/multilingual scores, and top open-model leaderboard placements on release. 289.8k Hugging Face downloads, not gated. Weights use the Qwen License — custom, non-OSI (acceptable-use policy, attribution) — describe as "open weights", never "open source". Superseded by the Qwen3 family, but still one of the most-deployed 70B-class models and a common vLLM/Ollama choice.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported (Alibaba, Sep 2024): competitive with Llama 3.1-70B on MMLU, HumanEval, MATH; strong multilingual — provider-reported. Independently re-benchmarked extensively; treat 2024 comparisons as historical.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 7/10
Scored 7/10 (verified tier, metrics-v1). Evidence: HF downloads 289,796 (peer-normalized); released 757 days ago. Manual review: Qwen2.5's 72B flagship (128K, 29 langs, 2024 leaderboard-topper); superseded by Qwen3. License: Qwen License — custom, non-OSI (acceptable-use policy); describe as 'open weights', never 'open source'.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 8.1/10
- Maintenance · recency, releases, issue load
- 4.0/10
- Openness · license + docs
- 7.2/10
- Manual review · sanity read + red-flag check
- 8.0/10
- · HF downloads 289,796 (peer-normalized)
- · released 757 days ago
- · manual review: Qwen2.5's 72B flagship (128K, 29 langs, 2024 leaderboard-topper); superseded by Qwen3