Model
Model explorer

CodeQwen1.5-7B

OPEN
Alibaba · Qwen1.5 family · released Apr 16, 2024

First Qwen code specialist: 3T code tokens, 92 languages, 64K context, HumanEval 83.5 (Chat).

ReasoningCodingVisionFunction callingTool useAgentic
526.5
Elo · unrated
Parameters
7B
Active params
7B (dense)
Context
64K tokens
Architecture
Dense decoder-only Transformer (GQA), code-specialized
License
Tongyi Qianwen License Agreement
Languages
—
API price (in/out)
No hosted API
Modalities
text
Benchmark results
Bar shows position within the tracked field; marker = field best
BigCodeBenchCoding45.6%#5
best: GPT-4o mini · 57.4%
HumanEval+Coding45.7%#30
best: Mistral Small 3.2 (24B) · 92.9%
HumanEvalCoding51.8%#118
best: Claude Opus 4.5 · 99.4%
LiveCodeBenchCoding7.9%#168
best: DeepSeek-V4-Pro (Think Max) · 93.5%
MBPP+Coding60.2%#28
best: Llama 3.1 405B · 88.6%
MBPPCoding72.2%#32
best: Llama-3.3-Nemotron-Super-49B v1 (Reasoning On) · 91.3%
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
Qwen1.5 family
Elo progression across releases
Qwen1.5-72BFeb 2024545.1Qwen1.5-MoE-A2.7BMar 202494.1CodeQwen1.5-7BApr 2024526.5Qwen1.5-110BApr 2024665.5
Resources
HFQwen/CodeQwen1.5-7B-ChatDOCSqwenlm.github.io/blog/codeqwen1.5
API price weights · each benchmark row carries its own source badge (see methodology)