Gemini 3.1 Pro Preview vs Gemini 3.5 Flash
| Spec | Gemini 3.1 Pro Preview | Gemini 3.5 Flash |
|---|---|---|
| Respan Index | 43.6 (#61) | 44.3 (#59) |
| Blended price per 1M | $4.5 | $3.38 |
| Context window | 1.0M | 1.0M |
| Input / output per 1M | $2 / $12 | $1.5 / $9 |
| Max output | 66K | 66K |
| Input | text, vision, audio | text, vision, audio |
| Released | 2026-02-19 | 2026-05-19 |
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 3.1 Pro Preview | Gemini 3.5 Flash |
|---|---|---|
| Coding | 14.7 Gemini 3.1 Pro Preview | 19.0 Gemini 3.5 Flash |
| Agents | 53.4 Gemini 3.1 Pro Preview | 49.7 Gemini 3.5 Flash |
| Reasoning | 45.8 Gemini 3.1 Pro Preview | 49.9 Gemini 3.5 Flash |
| Knowledge & language | 72.5 Gemini 3.1 Pro Preview | 65.6 Gemini 3.5 Flash |
| Human preference | 52.5 Gemini 3.1 Pro Preview | 53.8 Gemini 3.5 Flash |
Benchmarks both report
Gemini 3.1 Pro Preview leads on 16 of the 29 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | Gemini 3.1 Pro Preview | Gemini 3.5 Flash |
|---|---|---|
| LMArena WebDev (LMArena) | 1446 | 1500 |
| LMArena Elo | 1487 | 1479 |
| LMArena Vision (LMArena) | 1280 | 1282 |
| APEX-Agents (Mercor) | 35.1% | 27.5% |
| Chess Puzzles (Epoch AI) | 51% | 45% |
| DeepSWE v1.1 (Datacurve) | 11.7% | 36.1% |
| EBR-bench (Epoch AI) | 14.3% | 4.8% |
| FrontierMath Tier 4 v2 (Epoch AI) | 26.8% | 26.8% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 59.6% | 62.8% |
| GPQA Diamond | 94.3% | 92.8% |
| LiveBench Agentic Coding (LiveBench) | 44.1% | 49% |
| LiveBench Coding (LiveBench) | 76.5% | 78.2% |
| LiveBench Data Analysis (LiveBench) | 78.5% | 64.9% |
| LiveBench Instruction Following (LiveBench) | 79.1% | 75.6% |
| LiveBench Language (LiveBench) | 85.4% | 84.6% |
| LiveBench Mathematics (LiveBench) | 91% | 88.2% |
| LiveBench Reasoning (LiveBench) | 84% | 82% |
| Mystery Game Puzzles (Epoch AI) | 34% | 32% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 95.6% | 95.6% |
| SAGE (Vals AI) | 48.7% | 49.9% |
| SimpleQA Verified | 73.5% | 66.2% |
| SWE-bench Verified (Epoch AI) | 75.6% | 79.3% |
| Terminal-Bench 4.0 (Vals AI) | 2.5% | 6.1% |
| Vending-Bench 2 (Andon Labs) | 3774.25 | 5396.42 |
| WeirdML (Håvard Tveit Ihle) | 72.1% | 62.6% |
| AIME 2026 | 98.3% | 95% |
| Toolathlon-Verified (HKUST) | 61.1% | 67.3% |
| ARC-AGI-2 | 77.1% | 72.1% |
| SimpleBench (SimpleBench) | 79.6% | 76.7% |
FAQ
Is Gemini 3.1 Pro Preview better than Gemini 3.5 Flash?
Gemini 3.5 Flash scores 44.3 on the Respan Index (#59), ahead of Gemini 3.1 Pro Preview at 43.6 (#61). Gemini 3.1 Pro Preview leads on 16 of the 29 benchmarks both report.
Which is cheaper, Gemini 3.1 Pro Preview or Gemini 3.5 Flash?
Gemini 3.1 Pro Preview lists at $2 input and $12 output per 1M tokens; Gemini 3.5 Flash at $1.5 and $9. Gemini 3.5 Flash is 1.3x cheaper per token at list price.
Which has the larger context window?
Both accept 1.0M tokens of context.
Compare other models
- vsGemini 3.1 Pro Preview vs Gemini 3 Pro Preview
- vsGemini 3.5 Flash vs Gemini 2.5 Flash
- vsGemini 3.6 Flash vs Gemini 3.5 Flash
- vsClaude Sonnet 5.5 vs Gemini 3.1 Pro Preview
- vsGPT-5.6 Luna vs Gemini 3.1 Pro Preview
- vsGemini 3.1 Pro Preview vs Mistral Large 4
- vsGemini 3.1 Pro Preview vs GLM-5.2
- vsGemini 3.1 Pro Preview vs Qwen3.7-Max
- vsGemini 3.5 Flash vs Qwen3.8-Flash-Next
- vsGemini 3.5 Flash vs Mistral Large 4
- vsGemini 3.5 Flash vs Claude Opus 4.6