Models, measured side by side
Gemini 2.5 Pro vs Phi-4
Compare Gemini 2.5 Pro, Phi-4 by token pricing, context length, benchmark results, speed and tool support.
Monthly workload
Monthly cost breakdown
InputOutput
Gemini 2.5 Pro
$75 / mo
Phi-4
$2.10 / mo
Performance comparison
Gemini 2.5 Pro
16.1
Phi-4
5.9
Overview
Pricing & capacity
| Pricing & capacity | Gemini 2.5 Pro | Phi-4 |
|---|---|---|
| Input / 1M tokens | $1.25 | $0.07 |
| Output / 1M tokens | $10 | $0.14 |
| Cached input / 1M | $0.13 | Not listed |
| Monthly cost | $75 | $2.10 |
| Context window | 1M | 16K |
| Maximum output | 66K | 15K |
Input / 1M tokens
- Gemini 2.5 Pro
- $1.25
- Phi-4
- $0.07
Output / 1M tokens
- Gemini 2.5 Pro
- $10
- Phi-4
- $0.14
Cached input / 1M
- Gemini 2.5 Pro
- $0.13
- Phi-4
- Not listed
Monthly cost
- Gemini 2.5 Pro
- $75
- Phi-4
- $2.10
Context window
- Gemini 2.5 Pro
- 1M
- Phi-4
- 16K
Maximum output
- Gemini 2.5 Pro
- 66K
- Phi-4
- 15K
Benchmarks & performance
| Benchmarks & performance | Gemini 2.5 Pro | Phi-4 |
|---|---|---|
| Arena Elo | 1,255 | 1,123 |
| MMLU-Pro | 82.8% | 60.7% |
| Intelligence Index | 16.1 | 5.9 |
| GPQA | 84.4% | 57.5% |
| Humanity’s Last Exam | 22.5% | 3.8% |
| SciCode | 46.3% | Not listed |
| Output speed | 128 tokens/s | 39 tokens/s |
| Time to first token | 18.08 s | 2.59 s |
| DesignArena rating | 1,156 | Not listed |
| DesignArena win rate | 57.5% | Not listed |
Arena Elo
- Gemini 2.5 Pro
- 1,255
- Phi-4
- 1,123
MMLU-Pro
- Gemini 2.5 Pro
- 82.8%
- Phi-4
- 60.7%
Intelligence Index
- Gemini 2.5 Pro
- 16.1
- Phi-4
- 5.9
GPQA
- Gemini 2.5 Pro
- 84.4%
- Phi-4
- 57.5%
Humanity’s Last Exam
- Gemini 2.5 Pro
- 22.5%
- Phi-4
- 3.8%
SciCode
- Gemini 2.5 Pro
- 46.3%
- Phi-4
- Not listed
Output speed
- Gemini 2.5 Pro
- 128 tokens/s
- Phi-4
- 39 tokens/s
Time to first token
- Gemini 2.5 Pro
- 18.08 s
- Phi-4
- 2.59 s
DesignArena rating
- Gemini 2.5 Pro
- 1,156
- Phi-4
- Not listed
DesignArena win rate
- Gemini 2.5 Pro
- 57.5%
- Phi-4
- Not listed
Tools & features
| Tools & features | Gemini 2.5 Pro | Phi-4 |
|---|---|---|
| Function calling | Supported | Not listed |
| Structured outputs | Supported | Supported |
| JSON mode | Supported | Supported |
| Reasoning | Supported | Not listed |
| Deterministic seed | Supported | Supported |
| Prompt caching | Supported | Not listed |
Function calling
- Gemini 2.5 Pro
- Supported
- Phi-4
- Not listed
Structured outputs
- Gemini 2.5 Pro
- Supported
- Phi-4
- Supported
JSON mode
- Gemini 2.5 Pro
- Supported
- Phi-4
- Supported
Reasoning
- Gemini 2.5 Pro
- Supported
- Phi-4
- Not listed
Deterministic seed
- Gemini 2.5 Pro
- Supported
- Phi-4
- Supported
Prompt caching
- Gemini 2.5 Pro
- Supported
- Phi-4
- Not listed
API & availability
| API & availability | Gemini 2.5 Pro | Phi-4 |
|---|---|---|
| API identifier | gemini-2.5-pro | phi-4 |
| API providers | Google, Google, Google AI Studio, Google, Google AI Studio, Google, Google AI Studio | DeepInfra |
| Knowledge cutoff | 2025-01-31 | 2024-06-30 |
| Prices checked | Oct 8, 2026 | Oct 8, 2026 |
API identifier
- Gemini 2.5 Pro
gemini-2.5-pro- Phi-4
phi-4
API providers
- Gemini 2.5 Pro
- Google, Google, Google AI Studio, Google, Google AI Studio, Google, Google AI Studio
- Phi-4
- DeepInfra
Knowledge cutoff
- Gemini 2.5 Pro
- 2025-01-31
- Phi-4
- 2024-06-30
Prices checked
- Gemini 2.5 Pro
- Oct 8, 2026
- Phi-4
- Oct 8, 2026
About the models
Comparison FAQ
Gemini 2.5 Pro: $1.25 input and $10 output per million tokens. Phi-4: $0.07 input and $0.14 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.