Compare · ModelsLive · 2 picked · head to head
DeepSeek V4 Flash 0731 vs Gemini 3.6 Flash
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Gemini 3.6 Flash wins on 8/14 benchmarks
Gemini 3.6 Flash wins 8 of 14 shared benchmarks. Leads in reasoning · knowledge · general.
Category leads
reasoning·Gemini 3.6 Flashknowledge·Gemini 3.6 Flashgeneral·Gemini 3.6 Flashcoding·Gemini 3.6 Flashmath·DeepSeek V4 Flash 0731
Hype vs Reality
Attention vs performance
DeepSeek V4 Flash 0731
#80 by perf·#7 by attention
Gemini 3.6 Flash
#95 by perf·#6 by attention
Best value
DeepSeek V4 Flash 0731
7.1x better value than Gemini 3.6 Flash
DeepSeek V4 Flash 0731
168.6 pts/$
$0.33/M
Gemini 3.6 Flash
23.9 pts/$
$2.25/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
Google DeepMind
$4.27T·Tier 1
Head to head
14 benchmarks · 2 models
DeepSeek V4 Flash 0731Gemini 3.6 Flash
ARC-AGI
Gemini 3.6 Flash leads by +2.2
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
Gemini 3.6 Flash
91.2
ARC-AGI-2
DeepSeek V4 Flash 0731 leads by +1.0
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
Gemini 3.6 Flash
60.4
Chess Puzzles
Gemini 3.6 Flash leads by +10.5
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
DeepSeek V4 Flash 0731
28.4
Gemini 3.6 Flash
39.0
Dtbench
Gemini 3.6 Flash leads by +7.6
DeepSeek V4 Flash 0731
84.9
Gemini 3.6 Flash
92.5
Frontiercode
Gemini 3.6 Flash leads by +15.6
DeepSeek V4 Flash 0731
18.8
Gemini 3.6 Flash
34.4
FrontierMath-Tier-4-v2-Private
DeepSeek V4 Flash 0731 leads by +2.4
DeepSeek V4 Flash 0731
24.4
Gemini 3.6 Flash
21.9
FrontierMath-Tiers-1-3-v2-Private
Gemini 3.6 Flash leads by +1.4
DeepSeek V4 Flash 0731
57.5
Gemini 3.6 Flash
59.0
GPQA diamond
Gemini 3.6 Flash leads by +4.1
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
Gemini 3.6 Flash
92.2
Lmca
Gemini 3.6 Flash leads by +3.8
DeepSeek V4 Flash 0731
49.1
Gemini 3.6 Flash
52.9
Mystery Game Puzzles
DeepSeek V4 Flash 0731 leads by +4.4
DeepSeek V4 Flash 0731
27.3
Gemini 3.6 Flash
22.9
OTIS Mock AIME 2024-2025
DeepSeek V4 Flash 0731 leads by +0.3
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek V4 Flash 0731
94.4
Gemini 3.6 Flash
94.2
Proofbench
DeepSeek V4 Flash 0731 leads by +20.0
DeepSeek V4 Flash 0731
56.0
Gemini 3.6 Flash
36.0
SimpleQA Verified
Gemini 3.6 Flash leads by +32.6
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
Gemini 3.6 Flash
66.2
WeirdML
DeepSeek V4 Flash 0731 leads by +6.9
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
DeepSeek V4 Flash 0731
63.0
Gemini 3.6 Flash
56.1
Full benchmark table
| Benchmark | DeepSeek V4 Flash 0731 | Gemini 3.6 Flash |
|---|---|---|
ARC-AGI ARC-AGI · the original Abstraction and Reasoning Corpus, testing whether AI can solve novel visual pattern recognition tasks without memorization. | 89.0 | 91.2 |
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 | 60.4 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 28.4 | 39.0 |
Dtbench | 84.9 | 92.5 |
Frontiercode | 18.8 | 34.4 |
FrontierMath-Tier-4-v2-Private | 24.4 | 21.9 |
FrontierMath-Tiers-1-3-v2-Private | 57.5 | 59.0 |
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 | 92.2 |
Lmca | 49.1 | 52.9 |
Mystery Game Puzzles | 27.3 | 22.9 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 94.4 | 94.2 |
Proofbench | 56.0 | 36.0 |
SimpleQA Verified SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information. | 33.6 | 66.2 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 63.0 | 56.1 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.02 | $0.64 | 1.0M tokens (~524 books) | $1.72 | |
| $0.75 | $3.75 | 1.0M tokens (~524 books) | $15.00 |