Qwen 2.5
LatestAlibaba
•
Open Source•
72B# 58
Released
Sep 19, 2024
# 6
Knowledge Cutoff
Sep 24
# 9
Context Length
128K
Benchmarks
# 72
Code RankedAGI
38.0%
# 31
SWEBench Verified
23.8%
# 34
LiveCodeBench v5
31.1%
# 38
Code LMArena
1247
# 20
Code LiveBench (old)
57.6%
# 59
GPQA Diamond
49.0%
# 35
Reason LiveBench (old)
46.0%
# 42
ELO LMArena
1259
# 26
Math LiveBench (old)
52.4%
# 13
MATH
83.1%
# 19
Human Eval
86.6%
# 24
Human Eval+
51.2%
# 28
NYT Connections
11.1%
# 27
MMLU Pro
71.1%
# 16
MMLU
86.8%
# 27
Halluc. Hughes
4.7%
# 42
IF LiveBench (old)
64.4%
# 37
Avg LiveBench (old)
51.4%
Pricing
# 17
Input Cost /M
$0.38
# 16
Output Cost /M
$0.57