If you are choosing a Gemini 3-series model today, the practical shortlist is not Gemini 3 Pro versus Gemini 3.1 Pro. It is Gemini 3.1 Pro Preview, Gemini 3.1 Flash-Lite Preview, and Gemini 3 Flash Preview. Those are the three models most teams are realistically deciding between when they need to balance quality, speed, and cost.
Gemini 3 Pro Preview is no longer the real decision point because Google has already deprecated and shut it down. So this guide focuses on the three live options that still matter.
Quick answer
- Choose Gemini 3.1 Pro Preview if output quality matters most.
- Choose Gemini 3.1 Flash-Lite Preview if token cost matters most.
- Choose Gemini 3 Flash Preview if you want a stronger benchmark profile than Flash-Lite without paying Pro prices.
Gemini 3.1 Pro vs Gemini 3.1 Flash-Lite vs Gemini 3 Flash
| Area | Gemini 3.1 Pro Preview | Gemini 3.1 Flash-Lite Preview | Gemini 3 Flash Preview |
|---|---|---|---|
| Model ID | gemini-3.1-pro-preview | gemini-3.1-flash-lite-preview | gemini-3-flash-preview |
| Main role | Highest-quality flagship model | Lower-cost, high-volume model | Fast general-purpose model |
| Input tokens | 1,048,576 | 1,048,576 | 1,048,576 |
| Output tokens | 65,536 | 65,536 | 65,536 |
| Input price | $2 per 1M tokens up to 200k prompt tokens, then $4 | $0.25 per 1M text, image, or video tokens; $0.50 for audio | $0.50 per 1M text, image, or video tokens; $1 for audio |
| Output price | $12 per 1M output tokens up to 200k prompt tokens, then $18 | $1.50 per 1M output tokens | $3 per 1M output tokens |
| Best fit | Reasoning-heavy, premium tasks | Budget-sensitive, high-volume tasks | Better quality-speed balance below Pro |
How to think about the choice
The easiest way to think about these three models is to separate them by job.
- Gemini 3.1 Pro Preview is the quality-first choice.
- Gemini 3.1 Flash-Lite Preview is the cost-first choice.
- Gemini 3 Flash Preview sits in between as the more capable non-Pro option in the benchmark data we found.
That is why this is not really a version-number comparison. It is a model-tier decision. Flash-Lite is not simply "the newer Flash." It is a cheaper tier with a different job.
Pricing
If your main constraint is budget, the pricing table already narrows the decision quickly.
| Model | Input price | Output price | Pricing takeaway |
|---|---|---|---|
gemini-3.1-pro-preview | $2 per 1M tokens up to 200k prompt tokens, then $4 | $12 per 1M output tokens up to 200k prompt tokens, then $18 | Most expensive, justified only when you need top quality. |
gemini-3.1-flash-lite-preview | $0.25 per 1M text, image, or video tokens; $0.50 for audio | $1.50 per 1M output tokens | Cheapest option in this comparison. |
gemini-3-flash-preview | $0.50 per 1M text, image, or video tokens; $1 for audio | $3 per 1M output tokens | Costs about twice as much as Flash-Lite, but still far below Pro. |
If you only care about lowering token cost, pick Gemini 3.1 Flash-Lite. If you can afford more and want better model quality, then Gemini 3 Flash and Gemini 3.1 Pro are the more relevant options.
Token limits
All three models support a 1,048,576-token input window and a 65,536-token output limit in the Gemini API docs. That means the decision is not really about context-window size here. It is about capability, speed, and price.
| Model | Max input | Max output | Input types |
|---|---|---|---|
gemini-3.1-pro-preview | 1,048,576 | 65,536 | Text, image, video, audio, PDF |
gemini-3.1-flash-lite-preview | 1,048,576 | 65,536 | Text, image, video, audio, PDF |
gemini-3-flash-preview | 1,048,576 | 65,536 | Text, image, video, audio, PDF |
Official benchmarks
The cleanest benchmark view is to use only the metrics that appear across all three models in the official Google benchmark materials we reviewed. That gives a fairer shortlist table, even though the Flash and Flash-Lite figures still come from separate official Google pages rather than one single vendor comparison page.
| Official benchmark | Gemini 3.1 Pro | Gemini 3 Flash | Gemini 3.1 Flash-Lite |
|---|---|---|---|
| Humanity's Last Exam, no tools | 44.4% | 33.7% | 16.0% |
| GPQA Diamond, no tools | 94.3% | 90.4% | 86.9% |
| MMMU-Pro | 81.0% | 81.2% | 76.8% |
| Terminal / agent-style benchmark | 68.5% on Terminal-Bench 2.0 | 47.6% on Terminal-Bench 2.0 | Not available in the same benchmark family |
The first three rows are the most useful common signals across the shortlist. On those, Gemini 3.1 Pro leads overall, Gemini 3 Flash usually lands in the middle, and Gemini 3.1 Flash-Lite is the cheaper but weaker option on quality.
Important caveat: Google does not appear to publish one official single-table head-to-head comparison for all three of these models. So this shortlist table is compiled from official Google benchmark sources, using only the most comparable metrics we could find.
Which model should you choose?
| If you need... | Pick this model | Why |
|---|---|---|
| Best overall quality | Gemini 3.1 Pro Preview | Best official reasoning profile in this shortlist. |
| Lowest token cost | Gemini 3.1 Flash-Lite Preview | Cheapest model here by a clear margin. |
| Better benchmark quality than Flash-Lite without paying Pro rates | Gemini 3 Flash Preview | Stronger benchmark profile than Flash-Lite in the official compiled crosswalk. |
If you want the shortest possible decision rule, it is this: pick Gemini 3.1 Pro for quality, Gemini 3.1 Flash-Lite for cost, and Gemini 3 Flash for the middle ground.
Verdict
For most real model-picking decisions today, Gemini 3 Pro no longer matters. The real choice is between Gemini 3.1 Pro, Gemini 3.1 Flash-Lite, and Gemini 3 Flash. Among those three, Gemini 3.1 Pro is the premium quality option, Gemini 3.1 Flash-Lite is the budget option, and Gemini 3 Flash is the more capable non-Pro middle lane.
Frequently Asked Questions
Should I still compare Gemini 3 Pro when picking a model today?
Usually no. Google has deprecated and shut down Gemini 3 Pro Preview, so the more practical shortlist is Gemini 3.1 Pro, Gemini 3.1 Flash-Lite, and Gemini 3 Flash.
Is Gemini 3.1 Flash-Lite better than Gemini 3 Flash?
Not on the compiled official benchmark crosswalk used here. Gemini 3 Flash usually looks stronger on quality, while Gemini 3.1 Flash-Lite wins on lower cost.
Which Gemini model is the best value?
That depends on your workload. Gemini 3.1 Flash-Lite is the cheapest, Gemini 3.1 Pro is the strongest, and Gemini 3 Flash is the middle option if you want better benchmark quality than Flash-Lite without paying Pro pricing.
Sources
- Gemini API model list
- Gemini API deprecations
- Gemini 3 Flash Preview model page
- Gemini 3.1 Pro Preview model page
- Gemini 3.1 Flash-Lite Preview model page
- Gemini Developer API pricing
- Google DeepMind Gemini benchmarks
- Google DeepMind Gemini 3 Flash benchmarks
- Google DeepMind Gemini 3.1 Flash-Lite benchmarks
- Gemini 3.1 Flash-Lite model card
- Gemini 3.1 Pro announcement