Compare · ModelsLive · 2 picked · head to head

Qwen3 235B A22B Thinking 2507 vs DeepSeek V3

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

Winner summary

Qwen3 235B A22B Thinking 2507 wins 7 of 7 shared benchmarks. Leads in arena · knowledge · math.

Category leads
arena·Qwen3 235B A22B Thinking 2507knowledge·Qwen3 235B A22B Thinking 2507math·Qwen3 235B A22B Thinking 2507coding·Qwen3 235B A22B Thinking 2507
Hype vs Reality
Qwen3 235B A22B Thinking 2507
#66 by perf·no signal
QUIET
DeepSeek V3
#45 by perf·no signal
QUIET
Best value
1.4x better value than Qwen3 235B A22B Thinking 2507
Qwen3 235B A22B Thinking 2507
68.0 pts/$
$0.82/M
DeepSeek V3
97.5 pts/$
$0.60/M
Vendor risk
One or more vendors flagged
Alibaba Qwen logo
Alibaba (Qwen)
$293.0B·Tier 1
Low risk
DeepSeek logo
DeepSeek
$3.4B·Tier 1
Higher risk
Head to head
Qwen3 235B A22B Thinking 2507DeepSeek V3
Chatbot Arena Elo · Overall
Qwen3 235B A22B Thinking 2507 leads by +41.6
Qwen3 235B A22B Thinking 2507
1399.8
DeepSeek V3
1358.2
Fiction.LiveBench
Qwen3 235B A22B Thinking 2507 leads by +25.0
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
Qwen3 235B A22B Thinking 2507
75.0
DeepSeek V3
50.0
FrontierMath-2025-02-28-Private
Qwen3 235B A22B Thinking 2507 leads by +6.8
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
Qwen3 235B A22B Thinking 2507
8.5
DeepSeek V3
1.7
GPQA diamond
Qwen3 235B A22B Thinking 2507 leads by +31.4
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Qwen3 235B A22B Thinking 2507
73.4
DeepSeek V3
42.0
Lech Mazur Writing
Qwen3 235B A22B Thinking 2507 leads by +8.0
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
Qwen3 235B A22B Thinking 2507
85.0
DeepSeek V3
77.0
OTIS Mock AIME 2024-2025
Qwen3 235B A22B Thinking 2507 leads by +70.9
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Qwen3 235B A22B Thinking 2507
86.7
DeepSeek V3
15.8
WeirdML
Qwen3 235B A22B Thinking 2507 leads by +5.0
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Qwen3 235B A22B Thinking 2507
41.0
DeepSeek V3
36.1
Full benchmark table
BenchmarkQwen3 235B A22B Thinking 2507DeepSeek V3
Chatbot Arena Elo · Overall
1399.81358.2
Fiction.LiveBench
Fiction.LiveBench · a continuously updated benchmark using recently published fiction to test reading comprehension and reasoning, preventing data contamination.
75.050.0
FrontierMath-2025-02-28-Private
FrontierMath (Feb 2025) · original research-level math problems created by mathematicians, testing capabilities at the boundary of current AI mathematical reasoning.
8.51.7
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
73.442.0
Lech Mazur Writing
Lech Mazur Writing · evaluates creative writing ability, assessing prose quality, narrative coherence, and stylistic sophistication.
85.077.0
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
86.715.8
WeirdML
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
41.036.1
Pricing · per 1M tokens · projected $/mo at 10M tokens
ModelInputOutputContextProjected $/mo
Alibaba Qwen logoQwen3 235B A22B Thinking 2507$0.15$1.50131K tokens (~66 books)$4.86
DeepSeek logoDeepSeek V3$0.32$0.89164K tokens (~82 books)$4.63