Model Details
Qwen2 7B Instruct
Quick answer: Qwen2 7B Instruct is Alibaba's June 2024 compact model scoring 82.3% GSM8K, 79.9% HumanEval, 70.6% MMLU, 44.1% MMLU-Pro, and 8.41 MT-Bench. Apache 2.0 — the 7B instruction-tuned model in the Qwen2 family.
At a Glance
Where Qwen2 7B Instruct leads
- 79.9% HumanEval — competitive coding at 7B scale
- 82.3% GSM8K — solid math
- 8.41 MT-Bench — strong multi-turn instruction following
- Apache 2.0 — fully open
- Compact 7B: easy deployment, low cost
Where it lags
- 44.1% MMLU-Pro — limited advanced academic coverage
- 70.6% MMLU — moderate knowledge breadth
- June 2024: significantly superseded by Qwen2.5 7B, Qwen3 series
Best for: Legacy Qwen2 ecosystem; compact coding at 7B; simple multi-turn chatbot deployments; baselines and comparisons.
What Qwen2 7B Instruct Is
Qwen2 7B Instruct is the June 2024 7B instruction model from Alibaba's Qwen2 generation. It was competitive at launch but is now superseded by Qwen2.5 7B (Sep 2024) and Qwen3 variants — both of which improve substantially on all metrics.
For new deployments, Qwen2.5 7B Instruct or Qwen3 7B are recommended over Qwen2 7B.
Specifications
| Field | Value |
|---|---|
| Organization | Alibaba |
| License | Apache 2.0 |
| HuggingFace | Qwen/Qwen2-7B-Instruct |
| Release date | June 7, 2024 |
| Parameters | 7B |
| Modality | Text only |
| Context window | 128K tokens |
Pricing
Open weights under Apache 2.0 — self-host at no cost.
Public Benchmark Scores
| Benchmark | Score | Source | Date |
|---|---|---|---|
| GSM8K | 82.3% | Benchgen evaluation | 2024-06 |
| HumanEval | 79.9% | Benchgen evaluation | 2024-06 |
| MMLU | 70.6% | Benchgen evaluation | 2024-06 |
| MMLU-Pro | 44.1% | Benchgen evaluation | 2024-06 |
| MT-Bench | 8.41 | Benchgen evaluation | 2024-06 |
Qwen2 7B vs Alternatives
| Model | GSM8K | HumanEval | MMLU-Pro | License |
|---|---|---|---|---|
| Qwen2 7B Instruct | 82.3% | 79.9% | 44.1% | Apache 2.0 |
| Qwen2.5 7B Instruct | 91.6% | 84.8% | 56.3% | Apache 2.0 |
| Qwen2.5 14B Instruct | 94.8% | 83.5% | 64.0% | Apache 2.0 |
Qwen2.5 7B (Sep 2024) uniformly outperforms Qwen2 7B. For all new 7B deployments: use Qwen2.5 7B or Qwen3 7B.
Related Models & Benchmarks
- Compare: Qwen2.5 7B Instruct · Qwen2.5 14B Instruct
- Benchmarks: GSM8K · HumanEval · MMLU
Frequently Asked Questions
What is Qwen2 7B Instruct?
Alibaba's June 2024 7B instruction model scoring 82.3% GSM8K, 79.9% HumanEval, 70.6% MMLU. Apache 2.0 — superseded by Qwen2.5 7B for new deployments.Specs from Alibaba's Qwen2 7B Instruct release (June 2024) and Benchgen evaluations. Last updated 2026-07-24.
Benchmark Leaderboards
This model isn’t on any benchmark leaderboard yet.