Compare · ModelsLive · 2 picked · head to head
DeepSeek V4 Pro vs Qwen3.7 Max
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Qwen3.7 Max wins on 14/22 benchmarks
Qwen3.7 Max wins 14 of 22 shared benchmarks. Leads in speed · general · math.
Category leads
speed·Qwen3.7 Maxknowledge·DeepSeek V4 Progeneral·Qwen3.7 Maxmath·Qwen3.7 Maxcoding·DeepSeek V4 Proreasoning·DeepSeek V4 Prolanguage·Qwen3.7 Max
Hype vs Reality
Attention vs performance
DeepSeek V4 Pro
#89 by perf·#6 by attention
Qwen3.7 Max
#41 by perf·#2 by attention
Best value
DeepSeek V4 Pro
7.0x better value than Qwen3.7 Max
DeepSeek V4 Pro
172.1 pts/$
$0.31/M
Qwen3.7 Max
24.5 pts/$
$2.50/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
22 benchmarks · 2 models
DeepSeek V4 ProQwen3.7 Max
Artificial Analysis · Agentic Index
DeepSeek V4 Pro leads by +5.8
Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?"
DeepSeek V4 Pro
36.4
Qwen3.7 Max
30.6
Artificial Analysis · Coding Index
Qwen3.7 Max leads by +6.6
Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads.
DeepSeek V4 Pro
59.4
Qwen3.7 Max
66.0
Artificial Analysis · Quality Index
Qwen3.7 Max leads by +10.0
DeepSeek V4 Pro
36.0
Qwen3.7 Max
46.0
Chess Puzzles
DeepSeek V4 Pro leads by +1.1
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
DeepSeek V4 Pro
15.8
Qwen3.7 Max
14.8
Dtbench
Qwen3.7 Max leads by +2.7
DeepSeek V4 Pro
84.5
Qwen3.7 Max
87.1
FrontierMath-Tier-4-v2-Private
Qwen3.7 Max leads by +31.7
DeepSeek V4 Pro
2.4
Qwen3.7 Max
34.1
FrontierMath-Tiers-1-3-v2-Private
Qwen3.7 Max leads by +19.3
DeepSeek V4 Pro
45.3
Qwen3.7 Max
64.6
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
DeepSeek V4 Pro
87.9
Qwen3.7 Max
87.9
LiveBench · Agentic Coding
DeepSeek V4 Pro leads by +5.0
DeepSeek V4 Pro
56.7
Qwen3.7 Max
51.7
LiveBench · Coding
Qwen3.7 Max leads by +4.2
DeepSeek V4 Pro
70.0
Qwen3.7 Max
74.2
LiveBench · Data Analysis
DeepSeek V4 Pro leads by +2.8
DeepSeek V4 Pro
74.5
Qwen3.7 Max
71.8
LiveBench · If
Qwen3.7 Max leads by +11.7
DeepSeek V4 Pro
62.4
Qwen3.7 Max
74.0
LiveBench · Language
Qwen3.7 Max leads by +1.6
DeepSeek V4 Pro
78.1
Qwen3.7 Max
79.7
LiveBench · Mathematics
DeepSeek V4 Pro leads by +5.4
DeepSeek V4 Pro
90.7
Qwen3.7 Max
85.3
LiveBench · Overall
Qwen3.7 Max leads by +0.7
DeepSeek V4 Pro
73.6
Qwen3.7 Max
74.3
LiveBench · Reasoning
Qwen3.7 Max leads by +0.7
DeepSeek V4 Pro
82.7
Qwen3.7 Max
83.3
Lmca
Qwen3.7 Max leads by +3.3
DeepSeek V4 Pro
48.5
Qwen3.7 Max
51.8
Mystery Game Puzzles
Qwen3.7 Max leads by +16.5
DeepSeek V4 Pro
8.6
Qwen3.7 Max
25.1
OTIS Mock AIME 2024-2025
DeepSeek V4 Pro leads by +1.1
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek V4 Pro
96.7
Qwen3.7 Max
95.5
Proofbench
Qwen3.7 Max leads by +10.0
DeepSeek V4 Pro
16.0
Qwen3.7 Max
26.0
SimpleQA Verified
Qwen3.7 Max leads by +8.8
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
DeepSeek V4 Pro
47.0
Qwen3.7 Max
55.8
SWE-Bench verified
DeepSeek V4 Pro leads by +0.4
SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability.
DeepSeek V4 Pro
77.6
Qwen3.7 Max
77.3
Full benchmark table
| Benchmark | DeepSeek V4 Pro | Qwen3.7 Max |
|---|---|---|
Artificial Analysis · Agentic Index Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?" | 36.4 | 30.6 |
Artificial Analysis · Coding Index Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads. | 59.4 | 66.0 |
Artificial Analysis · Quality Index | 36.0 | 46.0 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 15.8 | 14.8 |
Dtbench | 84.5 | 87.1 |
FrontierMath-Tier-4-v2-Private | 2.4 | 34.1 |
FrontierMath-Tiers-1-3-v2-Private | 45.3 | 64.6 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 87.9 | 87.9 |
LiveBench · Agentic Coding | 56.7 | 51.7 |
LiveBench · Coding | 70.0 | 74.2 |
LiveBench · Data Analysis | 74.5 | 71.8 |
LiveBench · If | 62.4 | 74.0 |
LiveBench · Language | 78.1 | 79.7 |
LiveBench · Mathematics | 90.7 | 85.3 |
LiveBench · Overall | 73.6 | 74.3 |
LiveBench · Reasoning | 82.7 | 83.3 |
Lmca | 48.5 | 51.8 |
Mystery Game Puzzles | 8.6 | 25.1 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 96.7 | 95.5 |
Proofbench | 16.0 | 26.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 47.0 | 55.8 |
SWE-Bench verified SWE-bench Verified · 500 human-validated tasks from 12 real Python repositories (Django, Flask, scikit-learn, sympy, and others). Each task requires the model to produce a git patch that resolves a real GitHub issue and passes the test suite. The verified subset eliminates ambiguous tasks from the original SWE-bench. Claude Mythos Preview leads at 93.9%, crossing 90% for the first time in 2026. Opus 4.6 scores 80.8%. The benchmark remains the most-cited evaluation for code-generation capability. | 77.6 | 77.3 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.21 | $0.42 | 1.0M tokens (~524 books) | $2.61 | |
| $1.25 | $3.75 | 1.0M tokens (~500 books) | $18.75 |