Compare · ModelsLive · 2 picked · head to head

DeepSeek V3 vs Qwen3 235B A22B

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

Winner summary

Qwen3 235B A22B wins 8 of 8 shared benchmarks. Leads in coding · arena · knowledge.

Category leads
coding·Qwen3 235B A22Barena·Qwen3 235B A22Bknowledge·Qwen3 235B A22Bmath·Qwen3 235B A22Breasoning·Qwen3 235B A22B
Hype vs Reality
DeepSeek V3
#45 by perf·no signal
QUIET
Qwen3 235B A22B
#60 by perf·no signal
QUIET
Best value
2.0x better value than Qwen3 235B A22B
DeepSeek V3
97.5 pts/$
$0.60/M
Qwen3 235B A22B
49.6 pts/$
$1.14/M
Vendor risk
One or more vendors flagged
DeepSeek logo
DeepSeek
$3.4B·Tier 1
Higher risk
Alibaba Qwen logo
Alibaba (Qwen)
$293.0B·Tier 1
Low risk
Head to head
DeepSeek V3Qwen3 235B A22B
Aider polyglot
Qwen3 235B A22B leads by +11.2
Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework.
DeepSeek V3
48.4
Qwen3 235B A22B
59.6
Chatbot Arena Elo · Overall
Qwen3 235B A22B leads by +16.3
DeepSeek V3
1358.2
Qwen3 235B A22B
1374.4
Fiction.LiveBench
Qwen3 235B A22B leads by +17.7
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
DeepSeek V3
50.0
Qwen3 235B A22B
67.7
GPQA diamond
Qwen3 235B A22B leads by +18.9
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
DeepSeek V3
42.0
Qwen3 235B A22B
60.9
Lech Mazur Writing
Qwen3 235B A22B leads by +6.0
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
DeepSeek V3
77.0
Qwen3 235B A22B
83.0
MATH level 5
Qwen3 235B A22B leads by +4.0
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
DeepSeek V3
64.8
Qwen3 235B A22B
68.9
SimpleBench
Qwen3 235B A22B leads by +14.5
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
DeepSeek V3
2.7
Qwen3 235B A22B
17.2
WeirdML
Qwen3 235B A22B leads by +1.2
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
DeepSeek V3
36.1
Qwen3 235B A22B
37.3
Full benchmark table
BenchmarkDeepSeek V3Qwen3 235B A22B
Aider polyglot
Aider Polyglot · measures how well AI models can edit code across multiple programming languages using the Aider coding assistant framework.
48.459.6
Chatbot Arena Elo · Overall
1358.21374.4
Fiction.LiveBench
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
50.067.7
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
42.060.9
Lech Mazur Writing
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
77.083.0
MATH level 5
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
64.868.9
SimpleBench
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
2.717.2
WeirdML
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
36.137.3
Pricing · per 1M tokens · projected $/mo at 10M tokens
ModelInputOutputContextProjected $/mo
DeepSeek logoDeepSeek V3$0.32$0.89164K tokens (~82 books)$4.63
Alibaba Qwen logoQwen3 235B A22B$0.46$1.82131K tokens (~66 books)$7.96