Model

Qwen2-7B

OPEN

Alibaba · Qwen2 family · released Jun 7, 2024

Widely-used small Gen-2 dense model; Apache 2.0, 128K context on the Instruct variant.

ReasoningCodingVisionFunction callingTool useAgentic

485.5

Elo · rank #309

Parameters

7.07B

Active params

7.07B (dense)

Context

128K tokens

Architecture

Dense decoder-only Transformer (GQA)

License

Apache 2.0

Languages

30+

API price (in/out)

No hosted API

Modalities

text

Benchmark results

Bar shows position within the tracked field; marker = field best

ARC-ChallengeReasoning60.6%#87

best: Llama 3.1 405B · 96.9%

BIG-Bench HardReasoning62.6%#76

best: ERNIE 4.5 300B-A47B · 94.3%

C-EvalKnowledge83.2%#29

best: Qwen3.6-Plus · 93.3%

CMMLUKnowledge83.9%#7

best: Doubao-1.5-Pro · 90.9%

Gaokao MathMath35.1%#8

best: Doubao-Seed-1.6 · 96.0%

GPQA DiamondReasoning31.8%#255

best: GPT-5.6 · 94.6%

GSM8KMath79.9%#82

best: Llama 3.1 405B · 96.8%

HellaSwagReasoning80.7%#76

best: Claude 3 Opus · 95.4%

HumanEvalCoding51.2%#120

best: Claude Opus 4.5 · 99.4%

IFEvalReasoning54.7%#129

best: Gemma 4 26B A4B · 98.5%

KoMT-BenchHuman preference7.69#4

best: LG EXAONE 3.0 7.8B Instruct · 8.92

LiveCodeBenchCoding26.6%#157

best: DeepSeek-V4-Pro (Think Max) · 93.5%

LogicKorHuman preference6.3#25

best: GPT-4o · 9.33

MATH-500Math44.2%#143

best: GPT-5 · 99.4%

MBPPCoding65.9%#46

best: Llama-3.3-Nemotron-Super-49B v1 (Reasoning On) · 91.3%

MGSMMath57.0%#43

best: OpenAI o4-mini · 93.7%

MMLU (EU-21 languages)Knowledge58.4%#4

best: Llama 3.1 70B · 77.1%

MMLU-ProKnowledge40.0%#150

best: Claude Fable 5 · 91.5%

MMLUKnowledge70.3%#144

best: OpenAI o3 · 92.9%

MT-BenchHuman preference8.41#30

best: Hunyuan-Large (A52B) · 9.4

TruthfulQAKnowledge54.2%#40

best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%

WinoGrandeReasoning77.0%#52

best: PaLM 2 · 90.9%

Run it locally

VRAM @ Q4

5 GB

VRAM @ FP16

16 GB

Fits on (Q4)

RTX 3060 12GBRTX 4070 Ti 16GBRTX 3090 24GBRTX 4090 24GBRTX 5090 32GBM4 Pro 48GBM3 Max 128GBM3 Ultra 512GBA100 80GBH100 80GBH200 141GBB200 192GB

Throughput data unavailable.

Quantizations

GGUF · AWQ · GPTQ · MLX

Qwen2 family

Elo progression across releases

Qwen2-57B-A14BJun 2024589.2 Qwen2-72BJun 2024832.7 Qwen2-7BJun 2024485.5

Resources

HFQwen/Qwen2-7B-Instruct GHQwenLM/Qwen2 DOCSqwenlm.github.io/blog/qwen2

API price weights · each benchmark row carries its own source badge (see methodology)