Gemini 3 Pro Preview vs GPT-5
| Spec | Gemini 3 Pro Preview | GPT-5 |
|---|---|---|
| Respan Index | 39.7 (#69) | 39.5 (#71) |
| Blended price per 1M | - | $3.44 |
| Context window | 1.0M | 400K |
| Input / output per 1M | - | $1.25 / $10 |
| Max output | 66K | 128K |
| Input | text, vision, audio, video | text, vision |
| Released | 2025-11-18 | 2025-08-07 |
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | Gemini 3 Pro Preview | GPT-5 |
|---|---|---|
| Coding | 18.4 Gemini 3 Pro Preview | 10.9 GPT-5 |
| Agents | 46.9 Gemini 3 Pro Preview | 54.1 GPT-5 |
| Reasoning | 43.0 Gemini 3 Pro Preview | 42.2 GPT-5 |
| Knowledge & language | 47.5 Gemini 3 Pro Preview | 50.0 GPT-5 |
| Human preference | 52.4 Gemini 3 Pro Preview | 51.3 GPT-5 |
Benchmarks both report
Gemini 3 Pro Preview leads on 12 of the 15 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Gemini 3 Pro Preview | GPT-5 |
|---|---|---|
| LMArena Elo | 1486 | 1435 |
| LMArena Vision (LMArena) | 1289 | 1212 |
| LMArena WebDev (LMArena) | 1441 | 1418 |
| AIME 2025 | 95% | 94.6% |
| Chess Puzzles (Epoch AI) | 31% | 36% |
| GPQA Diamond | 91.9% | 86.2% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 91.4% | 91.4% |
| SAGE (Vals AI) | 47.6% | 43.7% |
| SWE-bench Verified (Epoch AI) | 72.9% | 73.6% |
| WeirdML (Håvard Tveit Ihle) | 69.9% | 60.7% |
| BALROG (BALROG) | 58.1% | 32.8% |
| GSO Opt@1 (GSO) | 17.7% | 5.9% |
| SimpleBench (SimpleBench) | 76.4% | 56.7% |
| SimpleQA Verified | 72.1% | 50.1% |
| SWE-bench Verified | 76.2% | 74.9% |
FAQ
Is Gemini 3 Pro Preview better than GPT-5?
Gemini 3 Pro Preview scores 39.7 on the Respan Index (#69), ahead of GPT-5 at 39.5 (#71). Gemini 3 Pro Preview leads on 12 of the 15 benchmarks both report.
Which has the larger context window?
Gemini 3 Pro Preview accepts 1.0M tokens and GPT-5 400K tokens.