Compare · ModelsLive · 2 picked · head to head

Qwen2.5 72B Instruct vs Qwen3.8 Max (0902)

Side by side · benchmarks, pricing, and signals you can act on.

CiteAdd another

Winner summary

Qwen3.8 Max (0902) wins on 5/5 benchmarks

Qwen3.8 Max (0902) wins 5 of 5 shared benchmarks. Leads in arena · general · knowledge.

Category leads

arena·Qwen3.8 Max (0902)general·Qwen3.8 Max (0902)knowledge·Qwen3.8 Max (0902)math·Qwen3.8 Max (0902)

Hype vs Reality

Attention vs performance

Qwen2.5 72B Instruct

#135 by perf·#2 by attention

DESERVED

Qwen3.8 Max (0902)

#85 by perf·#2 by attention

DESERVED

See full mindshare →

Best value

Qwen2.5 72B Instruct

9.5x better value than Qwen3.8 Max (0902)

Qwen2.5 72B Instruct

129.5 pts/$

$0.38/M

Qwen3.8 Max (0902)

13.7 pts/$

$4.00/M

Explore pricing →

Vendor risk

Who is behind the model

Alibaba (Qwen)

$293.0B·Tier 1

Low risk

Alibaba (Qwen)

$293.0B·Tier 1

Low risk

See the AI economy →

Head to head

5 benchmarks · 2 models

Qwen2.5 72B InstructQwen3.8 Max (0902)

Chatbot Arena Elo · Overall

Qwen3.8 Max (0902) leads by +178.8

Qwen2.5 72B Instruct

1302.8

Qwen3.8 Max (0902)

1481.6

Dtbench

Qwen3.8 Max (0902) leads by +48.5

Qwen2.5 72B Instruct

38.2

Qwen3.8 Max (0902)

86.7

GPQA diamond

Qwen3.8 Max (0902) leads by +58.0

Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.

Qwen2.5 72B Instruct

32.2

Qwen3.8 Max (0902)

90.2

Lmca

Qwen3.8 Max (0902) leads by +38.6

Qwen2.5 72B Instruct

15.8

Qwen3.8 Max (0902)

54.4

OTIS Mock AIME 2024-2025

Qwen3.8 Max (0902) leads by +91.5

OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.

Qwen2.5 72B Instruct

8.0

Qwen3.8 Max (0902)

99.4

Full benchmark table

Benchmark	Qwen2.5 72B Instruct	Qwen3.8 Max (0902)
Chatbot Arena Elo · Overall	1302.8	1481.6
Dtbench	38.2	86.7
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.	32.2	90.2
Lmca	15.8	54.4
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.	8.0	99.4

Pricing · per 1M tokens · projected $/mo at 10M tokens

Model	Input	Output	Context	Projected $/mo
Qwen2.5 72B Instruct	$0.36	$0.40	33K tokens (~16 books)	$3.70
Qwen3.8 Max (0902)	$2.00	$6.00	1.0M tokens (~500 books)	$30.00