Compare · ModelsLive · 2 picked · head to head
DeepSeek V3 vs GPT-4 Turbo
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
DeepSeek V3 wins on 8/13 benchmarks
DeepSeek V3 wins 8 of 13 shared benchmarks. Leads in reasoning · general · math.
Category leads
reasoning·DeepSeek V3general·DeepSeek V3knowledge·GPT-4 Turbomath·DeepSeek V3coding·DeepSeek V3
Hype vs Reality
Attention vs performance
DeepSeek V3
#70 by perf·no signal
GPT-4 Turbo
#197 by perf·no signal
Best value
DeepSeek V3
44.7x better value than GPT-4 Turbo
DeepSeek V3
87.2 pts/$
$0.64/M
GPT-4 Turbo
1.9 pts/$
$20.00/M
Vendor risk
Mixed exposure
One or more vendors flagged
DeepSeek
$3.4B·Tier 1
OpenAI
$840.0B·Tier 1
Head to head
13 benchmarks · 2 models
DeepSeek V3GPT-4 Turbo
BBH
DeepSeek V3 leads by +16.5
BIG-Bench Hard · a curated subset of 23 challenging tasks from BIG-Bench where language models previously failed to outperform average humans.
DeepSeek V3
83.3
GPT-4 Turbo
66.8
Dtbench
DeepSeek V3 leads by +3.5
DeepSeek V3
41.3
GPT-4 Turbo
37.8
GPQA diamond
DeepSeek V3 leads by +34.5
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
DeepSeek V3
42.0
GPT-4 Turbo
7.5
HellaSwag
GPT-4 Turbo leads by +8.5
HellaSwag · tests commonsense reasoning by asking models to predict the most plausible continuation of everyday scenarios.
DeepSeek V3
85.2
GPT-4 Turbo
93.7
Lmca
GPT-4 Turbo leads by +1.9
DeepSeek V3
18.2
GPT-4 Turbo
20.1
MATH level 5
DeepSeek V3 leads by +41.9
MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics.
DeepSeek V3
64.8
GPT-4 Turbo
23.0
Metr Time Horizons
DeepSeek V3 leads by +18.1
DeepSeek V3
47.4
GPT-4 Turbo
29.3
MMLU
DeepSeek V3 leads by +6.4
Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge.
DeepSeek V3
82.9
GPT-4 Turbo
76.5
OTIS Mock AIME 2024-2025
DeepSeek V3 leads by +14.7
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
DeepSeek V3
15.8
GPT-4 Turbo
1.0
SimpleBench
GPT-4 Turbo leads by +7.4
SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking.
DeepSeek V3
2.7
GPT-4 Turbo
10.1
TriviaQA
GPT-4 Turbo leads by +1.9
TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents.
DeepSeek V3
82.9
GPT-4 Turbo
84.8
WeirdML
DeepSeek V3 leads by +23.7
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
DeepSeek V3
36.1
GPT-4 Turbo
12.4
Winogrande
GPT-4 Turbo leads by +4.6
WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs.
DeepSeek V3
70.4
GPT-4 Turbo
75.0
Full benchmark table
| Benchmark | DeepSeek V3 | GPT-4 Turbo |
|---|---|---|
BBH BIG-Bench Hard · a curated subset of 23 challenging tasks from BIG-Bench where language models previously failed to outperform average humans. | 83.3 | 66.8 |
Dtbench | 41.3 | 37.8 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 42.0 | 7.5 |
HellaSwag HellaSwag · tests commonsense reasoning by asking models to predict the most plausible continuation of everyday scenarios. | 85.2 | 93.7 |
Lmca | 18.2 | 20.1 |
MATH level 5 MATH Level 5 · the hardest tier of the MATH benchmark, featuring competition-level problems from AMC, AIME, and Olympiad-style mathematics. | 64.8 | 23.0 |
Metr Time Horizons | 47.4 | 29.3 |
MMLU Massive Multitask Language Understanding · 57 subjects spanning STEM, humanities, social sciences, and more. The standard benchmark for broad knowledge. | 82.9 | 76.5 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 15.8 | 1.0 |
SimpleBench SimpleBench · tests fundamental reasoning capabilities with straightforward problems designed to expose gaps in basic logical and spatial thinking. | 2.7 | 10.1 |
TriviaQA TriviaQA · reading comprehension benchmark with trivia questions, requiring models to find and reason over evidence from provided documents. | 82.9 | 84.8 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 36.1 | 12.4 |
Winogrande WinoGrande · large-scale commonsense reasoning benchmark where models must resolve ambiguous pronouns in carefully constructed sentence pairs. | 70.4 | 75.0 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.26 | $1.03 | 164K tokens (~82 books) | $4.50 | |
| $10.00 | $30.00 | 128K tokens (~64 books) | $150.00 |
People also compared