Compare · ModelsLive · 3 picked · head to head
DeepSeek V4 Pro vs Kimi K2.7 Code vs Qwen3.7 Max
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Qwen3.7 Max wins on 15/32 benchmarks
Qwen3.7 Max wins 15 of 32 shared benchmarks. Leads in knowledge · math · reasoning.
Category leads
speed·DeepSeek V4 Proknowledge·Qwen3.7 Maxmath·Qwen3.7 Maxcoding·Kimi K2.7 Codereasoning·Qwen3.7 Maxlanguage·Qwen3.7 Maxgeneral·Qwen3.7 Max
Hype vs Reality
Attention vs performance
DeepSeek V4 Pro
#89 by perf·#6 by attention
Kimi K2.7 Code
#69 by perf·#17 by attention
Qwen3.7 Max
#41 by perf·#2 by attention
Best value
DeepSeek V4 Pro
5.6x better value than Kimi K2.7 Code
DeepSeek V4 Pro
172.1 pts/$
$0.31/M
Kimi K2.7 Code
30.5 pts/$
$1.84/M
Qwen3.7 Max
24.5 pts/$
$2.50/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
Moonshot AI
$18.0B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
32 benchmarks · 3 models
DeepSeek V4 ProKimi K2.7 CodeQwen3.7 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
Kimi K2.7 Code
29.6
Qwen3.7 Max
30.6
Artificial Analysis · Coding Index
Qwen3.7 Max leads by +5.2
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
Kimi K2.7 Code
60.8
Qwen3.7 Max
66.0
Artificial Analysis · Quality Index
Qwen3.7 Max leads by +10.0
DeepSeek V4 Pro
36.0
Kimi K2.7 Code
25.8
Qwen3.7 Max
46.0
Chess Puzzles
Kimi K2.7 Code 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
Kimi K2.7 Code
16.9
Qwen3.7 Max
14.8
FrontierMath-Tier-4-v2-Private
Qwen3.7 Max leads by +21.9
DeepSeek V4 Pro
2.4
Kimi K2.7 Code
12.2
Qwen3.7 Max
34.1
FrontierMath-Tiers-1-3-v2-Private
Qwen3.7 Max leads by +10.5
DeepSeek V4 Pro
45.3
Kimi K2.7 Code
54.0
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
Kimi K2.7 Code
83.8
Qwen3.7 Max
87.9
LiveBench · Agentic Coding
Kimi K2.7 Code leads by +13.3
DeepSeek V4 Pro
56.7
Kimi K2.7 Code
70.0
Qwen3.7 Max
51.7
LiveBench · Coding
Qwen3.7 Max leads by +0.3
DeepSeek V4 Pro
70.0
Kimi K2.7 Code
74.0
Qwen3.7 Max
74.2
LiveBench · Data Analysis
DeepSeek V4 Pro leads by +2.8
DeepSeek V4 Pro
74.5
Kimi K2.7 Code
62.7
Qwen3.7 Max
71.8
LiveBench · If
Qwen3.7 Max leads by +11.7
DeepSeek V4 Pro
62.4
Kimi K2.7 Code
56.3
Qwen3.7 Max
74.0
LiveBench · Language
Qwen3.7 Max leads by +1.6
DeepSeek V4 Pro
78.1
Kimi K2.7 Code
77.9
Qwen3.7 Max
79.7
LiveBench · Mathematics
DeepSeek V4 Pro leads by +5.4
DeepSeek V4 Pro
90.7
Kimi K2.7 Code
79.6
Qwen3.7 Max
85.3
LiveBench · Overall
Qwen3.7 Max leads by +0.7
DeepSeek V4 Pro
73.6
Kimi K2.7 Code
71.9
Qwen3.7 Max
74.3
LiveBench · Reasoning
Qwen3.7 Max leads by +0.5
DeepSeek V4 Pro
82.7
Kimi K2.7 Code
82.8
Qwen3.7 Max
83.3
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
Kimi K2.7 Code
95.5
Qwen3.7 Max
95.5
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
Kimi K2.7 Code
36.5
Qwen3.7 Max
55.8
Artificial Analysis · CritPt
DeepSeek V4 Pro leads by +8.0
DeepSeek V4 Pro
18.0
Kimi K2.7 Code
10.0
Artificial Analysis · GDPval
DeepSeek V4 Pro leads by +20.8
DeepSeek V4 Pro
47.1
Kimi K2.7 Code
26.3
Artificial Analysis · GPQA Diamond
DeepSeek V4 Pro leads by +3.2
DeepSeek V4 Pro
92.8
Kimi K2.7 Code
89.6
Artificial Analysis · Humanity's Last Exam
DeepSeek V4 Pro leads by +6.0
DeepSeek V4 Pro
41.0
Kimi K2.7 Code
35.0
Artificial Analysis · Long Context Reasoning
DeepSeek V4 Pro leads by +1.0
DeepSeek V4 Pro
80.3
Kimi K2.7 Code
79.3
Artificial Analysis · SciCode
DeepSeek V4 Pro leads by +3.2
DeepSeek V4 Pro
51.0
Kimi K2.7 Code
47.8
Dtbench
Qwen3.7 Max leads by +2.7
DeepSeek V4 Pro
84.5
Qwen3.7 Max
87.1
Frontiercode
Kimi K2.7 Code leads by +12.4
DeepSeek V4 Pro
17.6
Kimi K2.7 Code
30.1
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
Proofbench
Qwen3.7 Max leads by +10.0
DeepSeek V4 Pro
16.0
Qwen3.7 Max
26.0
SimpleBench
Qwen3.7 Max leads by +15.0
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
Kimi K2.7 Code
49.5
Qwen3.7 Max
64.5
Surface Evolver Bench
Kimi K2.7 Code leads by +8.8
DeepSeek V4 Pro
40.0
Kimi K2.7 Code
48.8
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
WeirdML
Kimi K2.7 Code leads by +5.2
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
DeepSeek V4 Pro
48.9
Kimi K2.7 Code
54.1
Full benchmark table
| Benchmark | DeepSeek V4 Pro | Kimi K2.7 Code | 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 | 29.6 | 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 | 60.8 | 66.0 |
Artificial Analysis · Quality Index | 36.0 | 25.8 | 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 | 16.9 | 14.8 |
FrontierMath-Tier-4-v2-Private | 2.4 | 12.2 | 34.1 |
FrontierMath-Tiers-1-3-v2-Private | 45.3 | 54.0 | 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 | 83.8 | 87.9 |
LiveBench · Agentic Coding | 56.7 | 70.0 | 51.7 |
LiveBench · Coding | 70.0 | 74.0 | 74.2 |
LiveBench · Data Analysis | 74.5 | 62.7 | 71.8 |
LiveBench · If | 62.4 | 56.3 | 74.0 |
LiveBench · Language | 78.1 | 77.9 | 79.7 |
LiveBench · Mathematics | 90.7 | 79.6 | 85.3 |
LiveBench · Overall | 73.6 | 71.9 | 74.3 |
LiveBench · Reasoning | 82.7 | 82.8 | 83.3 |
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 | 95.5 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 47.0 | 36.5 | 55.8 |
Artificial Analysis · CritPt | 18.0 | 10.0 | — |
Artificial Analysis · GDPval | 47.1 | 26.3 | — |
Artificial Analysis · GPQA Diamond | 92.8 | 89.6 | — |
Artificial Analysis · Humanity's Last Exam | 41.0 | 35.0 | — |
Artificial Analysis · Long Context Reasoning | 80.3 | 79.3 | — |
Artificial Analysis · SciCode | 51.0 | 47.8 | — |
Dtbench | 84.5 | — | 87.1 |
Frontiercode | 17.6 | 30.1 | — |
Lmca | 48.5 | — | 51.8 |
Mystery Game Puzzles | 8.6 | — | 25.1 |
Proofbench | 16.0 | — | 26.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | — | 49.5 | 64.5 |
Surface Evolver Bench | 40.0 | 48.8 | — |
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 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 48.9 | 54.1 | — |
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 | |
| $0.61 | $3.07 | 262K tokens (~131 books) | $12.26 | |
| $1.25 | $3.75 | 1.0M tokens (~500 books) | $18.75 |
People also compared