Qwen2.5-Coder 32B Instruct

Qwen2.5-Coder is a specialized coding model trained on 5.5 trillion tokens of code data, supporting 92 programming languages with a 128K context window. It excels in code generation, completion, repair, and multi-programming tasks while maintaining strong performance in mathematics and general capabilities.

HumanEval

92.7%

i
GSM8k

91.1%

i
MBPP

90.2%

i
HellaSwag

83.0%

i
Winogrande

80.8%

i
MMLU-Redux

77.5%

i
MMLU

75.1%

i
ARC-C

70.5%

i
MATH

57.2%

i
TruthfulQA

54.2%

i
MMLU-Pro

50.4%

i
BigCodeBench-Full

49.6%

i
TheoremQA

43.1%

i
LiveCodeBench

31.4%

i
BigCodeBench-Hard

27.0%

i

Pricing, uptime, and speed via OpenRouter — updated Jul 17, 2026, 04:19 AM.

Provider	Status	Input	Output	Limits	Uptime	Speed	Notes
Cloudflare	available	$0.66/Mtok	$1.00/Mtok	33K tokens context 33K tokens max output	—	481 ms p50 TTFT 16 tok/s p50