Compare · ModelsLive · 2 picked · head to head
DeepSeek V4 Flash 0731 vs o3
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek V4 Flash 0731 wins on 10/12 benchmarks
DeepSeek V4 Flash 0731 wins 10 of 12 shared benchmarks. Leads in reasoning · general · math.
Category leads
reasoning·DeepSeek V4 Flash 0731knowledge·o3general·DeepSeek V4 Flash 0731math·DeepSeek V4 Flash 0731coding·DeepSeek V4 Flash 0731
Hype vs Reality
Attention vs performance
DeepSeek V4 Flash 0731
#76 by perf·#6 by attention
o3
#111 by perf·no signal
Best value
DeepSeek V4 Flash 0731
8.4x better value than o3
DeepSeek V4 Flash 0731
85.5 pts/$
$0.65/M
o3
10.2 pts/$
$5.00/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
12 benchmarks · 2 models
DeepSeek V4 Flash 0731o3
ARC-AGI
DeepSeek V4 Flash 0731 leads by +28.2
ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization.
DeepSeek V4 Flash 0731
89.0
o3
60.8
ARC-AGI-2
DeepSeek V4 Flash 0731 leads by +54.9
ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data.
DeepSeek V4 Flash 0731
61.4
o3
6.5
Chess Puzzles
o3 leads by +5.3
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
DeepSeek V4 Flash 0731
29.5
o3
34.8
Dtbench
DeepSeek V4 Flash 0731 leads by +10.2
DeepSeek V4 Flash 0731
84.9
o3
74.7
FrontierMath-Tiers-1-3-v2-Private
DeepSeek V4 Flash 0731 leads by +24.2
DeepSeek V4 Flash 0731
57.5
o3
33.3
GPQA diamond
DeepSeek V4 Flash 0731 leads by +12.3
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
DeepSeek V4 Flash 0731
88.0
o3
75.8
Lmca
DeepSeek V4 Flash 0731 leads by +2.4
DeepSeek V4 Flash 0731
49.1
o3
46.7
Mystery Game Puzzles
DeepSeek V4 Flash 0731 leads by +5.5
DeepSeek V4 Flash 0731
27.3
o3
21.8
OTIS Mock AIME 2024-2025
DeepSeek V4 Flash 0731 leads by +10.0
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek V4 Flash 0731
94.4
o3
84.4
SimpleBench
DeepSeek V4 Flash 0731 leads by +9.6
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
DeepSeek V4 Flash 0731
53.3
o3
43.7
SimpleQA Verified
o3 leads by +15.8
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
DeepSeek V4 Flash 0731
33.6
o3
49.4
WeirdML
DeepSeek V4 Flash 0731 leads by +10.5
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
DeepSeek V4 Flash 0731
63.0
o3
52.4
Full benchmark table
| Benchmark | DeepSeek V4 Flash 0731 | o3 |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 89.0 | 60.8 |
ARC-AGI-2 ARC-AGI-2 · the second iteration of the Abstraction and Reasoning Corpus, testing novel pattern recognition and abstract reasoning without prior training data. | 61.4 | 6.5 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 29.5 | 34.8 |
Dtbench | 84.9 | 74.7 |
FrontierMath-Tiers-1-3-v2-Private | 57.5 | 33.3 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 88.0 | 75.8 |
Lmca | 49.1 | 46.7 |
Mystery Game Puzzles | 27.3 | 21.8 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 94.4 | 84.4 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 53.3 | 43.7 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 33.6 | 49.4 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 63.0 | 52.4 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.02 | $1.28 | 1.0M tokens (~524 books) | $3.31 | |
| $2.00 | $8.00 | 200K tokens (~100 books) | $35.00 |