Gemini 2.5 Pro vs Gemini 2.5 Pro Preview (May' 25)
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Gemini 2.5 Pro22.5
Gemini 2.5 Pro Preview (May' 25)19.9
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Gemini 2.5 Pro0
Gemini 2.5 Pro Preview (May' 25)—
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Gemini 2.5 Pro2.2
Gemini 2.5 Pro Preview (May' 25)—
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Gemini 2.5 Pro1,443
Gemini 2.5 Pro Preview (May' 25)—
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Gemini 2.5 Pro74.9
Gemini 2.5 Pro Preview (May' 25)—
| Benchmark | Gemini 2.5 Pro | Gemini 2.5 Pro Preview (May' 25) |
|---|---|---|
| Humanity's Last Exam | 22.5 | 19.9 |
| Terminal-Bench 4.0 | 0 | — |
| AutomationBench-AA | 2.2 | — |
| Arena Text · creative writing | 1,443 | — |
| MMMU-Pro (vision) | 74.9 | — |
Specs
| Spec | Gemini 2.5 Pro | Gemini 2.5 Pro Preview (May' 25) |
|---|---|---|
| Input, per 1M | $1.25 | — |
| Output, per 1M | $10 | — |
| Cost per task? | $0.33 | — |
| Speed, tok/s | 115 | — |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jun 2025 | May 2025 |
| See images | ✓ | — |
| Read PDFs | Partial | ? |
| Hear audio | ✓ | ✓ |
| Watch video | ✓ | ✓ |
| Call tools | ✓ | ? |
| Force a tool call | Partial | ? |
| Return strict JSON | ✓ | ? |
| Search the web | ✓ | ? |
| Run code | ✓ | ? |
| Use a computer | ? | ? |
What breaks when you switch
Gemini 2.5 Pro → Gemini 2.5 Pro Preview (May' 25)
No image input
Gemini 2.5 Pro Preview (May' 25) → Gemini 2.5 Pro
Nothing known.