Models, measured side by side
DeepSeek V3 vs Gemini 2.5 Pro
Compare DeepSeek V3, Gemini 2.5 Pro by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
DeepSeek V3
$10 / mo
Gemini 2.5 Pro
$75 / mo
Performance comparison
DeepSeek V3
8.5
Gemini 2.5 Pro
16.1
Overview
Pricing & capacity
| Pricing & capacity | DeepSeek V3 | Gemini 2.5 Pro |
|---|---|---|
| Input / 1M tokens | $0.26 | $1.25 |
| Output / 1M tokens | $1.03 | $10 |
| Cached input / 1M | Not listed | $0.13 |
| Monthly cost | $10 | $75 |
| Context window | 164K | 1M |
| Maximum output | 16K | 66K |
Input / 1M tokens
- DeepSeek V3
- $0.26
- Gemini 2.5 Pro
- $1.25
Output / 1M tokens
- DeepSeek V3
- $1.03
- Gemini 2.5 Pro
- $10
Cached input / 1M
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- $0.13
Monthly cost
- DeepSeek V3
- $10
- Gemini 2.5 Pro
- $75
Context window
- DeepSeek V3
- 164K
- Gemini 2.5 Pro
- 1M
Maximum output
- DeepSeek V3
- 16K
- Gemini 2.5 Pro
- 66K
Benchmarks & performance
| Benchmarks & performance | DeepSeek V3 | Gemini 2.5 Pro |
|---|---|---|
| Arena Elo | 1,211 | 1,255 |
| MMLU-Pro | 75.4% | 82.8% |
| Intelligence Index | 8.5 | 16.1 |
| GPQA | 55.7% | 84.4% |
| Humanity’s Last Exam | 2.9% | 22.5% |
| SciCode | 35.8% | 46.3% |
| Output speed | Not listed | 128 tokens/s |
| Time to first token | Not listed | 18.08 s |
| DesignArena rating | Not listed | 1,156 |
| DesignArena win rate | Not listed | 57.5% |
Arena Elo
- DeepSeek V3
- 1,211
- Gemini 2.5 Pro
- 1,255
MMLU-Pro
- DeepSeek V3
- 75.4%
- Gemini 2.5 Pro
- 82.8%
Intelligence Index
- DeepSeek V3
- 8.5
- Gemini 2.5 Pro
- 16.1
GPQA
- DeepSeek V3
- 55.7%
- Gemini 2.5 Pro
- 84.4%
Humanity’s Last Exam
- DeepSeek V3
- 2.9%
- Gemini 2.5 Pro
- 22.5%
SciCode
- DeepSeek V3
- 35.8%
- Gemini 2.5 Pro
- 46.3%
Output speed
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- 128 tokens/s
Time to first token
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- 18.08 s
DesignArena rating
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- 1,156
DesignArena win rate
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- 57.5%
Tools & features
| Tools & features | DeepSeek V3 | Gemini 2.5 Pro |
|---|---|---|
| Function calling | Supported | Supported |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Not listed | Supported |
| Deterministic seed | Supported | Supported |
| Prompt caching | Not listed | Supported |
Function calling
- DeepSeek V3
- Supported
- Gemini 2.5 Pro
- Supported
Structured outputs
- DeepSeek V3
- Supported
- Gemini 2.5 Pro
- Supported
JSON mode
- DeepSeek V3
- Supported
- Gemini 2.5 Pro
- Supported
Reasoning
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- Supported
Deterministic seed
- DeepSeek V3
- Supported
- Gemini 2.5 Pro
- Supported
Prompt caching
- DeepSeek V3
- Not listed
- Gemini 2.5 Pro
- Supported
API & availability
| API & availability | DeepSeek V3 | Gemini 2.5 Pro |
|---|---|---|
| API identifier | deepseek-v3 | gemini-2.5-pro |
| API providers | StreamLake, DeepInfra | Google, Google, Google AI Studio, Google, Google AI Studio, Google, Google AI Studio |
| Knowledge cutoff | 2024-07-31 | 2025-01-31 |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- DeepSeek V3
deepseek-v3- Gemini 2.5 Pro
gemini-2.5-pro
API providers
- DeepSeek V3
- StreamLake, DeepInfra
- Gemini 2.5 Pro
- Google, Google, Google AI Studio, Google, Google AI Studio, Google, Google AI Studio
Knowledge cutoff
- DeepSeek V3
- 2024-07-31
- Gemini 2.5 Pro
- 2025-01-31
Prices checked
- DeepSeek V3
- Oct 8, 2026
- Gemini 2.5 Pro
- Oct 8, 2026
About the models
Comparison FAQ
DeepSeek V3: $0.26 input and $1.03 output per million tokens. Gemini 2.5 Pro: $1.25 input and $10 output per million tokens. The cheaper choice depends on your input-to-output ratio; a model can have a lower input rate but a higher output rate.