Gemini 3 Flash Preview vs Gemini 3.1 Pro Preview
Both are Google models with identical 1,048,576-token context windows, so this comparison comes down to price and task fit, not context length. Flash Preview shipped December 17, 2025; Pro Preview followed February 19, 2026 with mandatory reasoning and a broader Design Arena benchmark spread. On a 10M-input/2M-output workload, Flash costs $11.00 and Pro Preview costs $44.00 on OpenKey — a 4x gap that should drive your default choice unless the task specifically needs Pro's edge.
Spec vs spec
| Spec | Gemini 3 Flash Preview | Gemini 3.1 Pro Preview |
|---|---|---|
| Context window | 1.0M | 1.0M |
| Max output | 66K | 66K |
| Input modalities | text, image, file, audio, video | audio, file, image, text, video |
| Output modalities | text | text |
| Released | Dec 17, 2025 | Feb 19, 2026 |
| Reasoning | optional | always on |
Pricing
Per 1M tokens. Provider price plus the flat 3% fee — the sum is what you pay.
google/gemini-3-flash-preview
Input · 1M tokens
$0.500 + 3%$0.515
Output · 1M tokens
$3.00 + 3%$3.09
Cache read · 1M tokens
$0.050 + 3%$0.052
Cache write · 1M tokens
$0.083 + 3%$0.086
FEE — FLAT, EVERY MODEL3%
google/gemini-3.1-pro-preview
Input · 1M tokens
$2.00 + 3%$2.06
Output · 1M tokens
$12.00 + 3%$12.36
Cache read · 1M tokens
$0.200 + 3%$0.206
Cache write · 1M tokens
$0.375 + 3%$0.386
FEE — FLAT, EVERY MODEL3%
One workload, priced on both
10M input + 2M output tokens at each model's price, flat 3% fee included.
google/gemini-3-flash-previewCheaper
$11.33
$11.00 provider + 3%
google/gemini-3.1-pro-preview
$45.32
$44.00 provider + 3%
Benchmarks
Design Arena categories where both models have results. Higher Elo and lower rank win.
| Gemini 3 Flash Preview | Gemini 3.1 Pro Preview | |||
|---|---|---|---|---|
| Category | Elo | Rank | Elo | Rank |
| 3D | 1261 | #29 | 1303 | #17 |
| agenticslides | 1073 | #9 | 1112 | #8 |
| agenticslides(python-pptx) | 1075 | #9 | 1107 | #8 |
| androidnative | 1062 | #24 | 1059 | #25 |
| Code | 1239 | #33 | 1290 | #18 |
| Full-stack | 1130 | #17 | 1139 | #16 |
| Game dev | 1232 | #38 | 1264 | #26 |
| godotgamedev | 1218 | #8 | 1220 | #6 |
| Mobile apps | 1188 | #17 | 1176 | #22 |
| Web apps | 1185 | #17 | 1196 | #13 |
| Websites | 1239 | #33 | 1294 | #15 |
Head-to-head preference voting. How we filter and rank
Pricing math
Provider list price for Gemini 3 Flash Preview is $0.50/M input and $3.00/M output. OpenKey adds its flat 3% fee: $0.50 × 1.03 = $0.515/M input, $3.00 × 1.03 = $3.09/M output. Gemini 3.1 Pro Preview lists at $2.00/M input and $12.00/M output, becoming $2.06/M and $12.36/M on OpenKey. That's a 4x ratio on input price (0.25 input_price_ratio) and roughly 4x on output.
Run the numbers on a 10M-input / 2M-output workload: Flash Preview costs $11.00 total, Pro Preview costs $44.00. If you're calling either model thousands of times a day, that gap compounds fast. Both run on OpenKey with one API key and the same flat 3% fee, so switching between them mid-project is a one-line change, not a new integration.
Agentic and coding benchmarks
On Design Arena's agents arena, the two trade places depending on category. Flash Preview ranks 9th in agenticslides (elo 1073, 39.3% win rate) while Pro Preview ranks 8th in the same category (elo 1112, 33.8% win rate) — close, with Pro nominally ahead on elo but Flash slightly ahead on win rate. In godotgamedev, Pro Preview ranks 6th (elo 1220, 53.6% win) versus Flash's 8th (elo 1218, 52.3% win) — nearly identical elo, Pro edges the rank.
Where Pro Preview clearly separates is in categories Flash doesn't compete in at all: SVG (rank 2, elo 1347, 70.3% win rate) and ASCII art (rank 4, elo 1314, 63.4% win rate). Pro Preview also reports Artificial Analysis scores — coding index 68.8, intelligence index 46.5, agentic index 21.4 — that Flash Preview's record doesn't include, so head-to-head on those specific metrics isn't possible from this data.
Context and modality
No difference here — both models support 1,048,576 tokens of context (context_ratio 1.0) and the same five input modalities: text, image, file, audio, video. Both output text only. Max completion tokens are effectively equal too: 65,535 for Flash Preview versus 65,536 for Pro Preview, a one-token rounding difference that doesn't matter in practice. If your use case is long-document analysis or multimodal ingestion, context and modality won't be the deciding factor — price and task-specific benchmark performance will be.
Reasoning behavior
Pro Preview makes reasoning mandatory, with effort levels of high, medium, or low (default: medium). Flash Preview makes reasoning optional and adds a minimal tier on top of high/medium/low (also defaulting to medium). If you're building a pipeline where you sometimes want fast, non-reasoning passes — classification, simple extraction, formatting — Flash Preview's flexibility to skip reasoning matters operationally, not just on price. Pro Preview always pays the reasoning-token tax.
Which model for which job
| Use case | Pick | Why |
|---|---|---|
| High-volume agent loops (tool calls, multi-turn chat) | Gemini 3 Flash Preview | $11.00 vs $44.00 on the same 10M-in/2M-out workload |
| SVG or vector graphic generation | Gemini 3.1 Pro Preview | Ranks 2nd in Design Arena's SVG category (elo 1347, 70.3% win rate) |
| UI component generation | Gemini 3.1 Pro Preview | Ranks 8th (elo 1322, 68.1% win rate) vs no comparable Flash entry |
| Simple, non-reasoning tasks (extraction, formatting) | Gemini 3 Flash Preview | Reasoning is optional, unlike Pro Preview's mandatory reasoning |
| Full-stack app scaffolding | Gemini 3 Flash Preview | Rank 17 (elo 1130) at a quarter of Pro Preview's input cost for a close 16th-place result |
| Cost-sensitive long-context batch jobs | Gemini 3 Flash Preview | Identical 1,048,576-token context at 4x lower price |
Questions
- How much cheaper is Gemini 3 Flash Preview than Gemini 3.1 Pro Preview?
- On OpenKey, Flash Preview runs $0.515/M input and $3.09/M output versus Pro Preview's $2.06/M input and $12.36/M output — a 4x gap on input (input_price_ratio 0.25). On a 10M-input/2M-output workload that's $11.00 versus $44.00.
- Do they have the same context window?
- Yes. Both support 1,048,576 tokens of context (context_ratio 1.0), so context length isn't a differentiator. Max output tokens are also nearly identical: 65,535 for Flash Preview and 65,536 for Pro Preview.
- Which one wins on agentic coding benchmarks?
- It's close and category-dependent. In godotgamedev, Pro Preview ranks 6th (elo 1220) versus Flash's 8th (elo 1218) — nearly tied. In agenticslides, Flash ranks 9th (elo 1073) versus Pro's 8th (elo 1112). Pro Preview also uniquely reports an Artificial Analysis coding index of 68.8.
- Is reasoning required on both models?
- No. Gemini 3.1 Pro Preview requires reasoning (mandatory: true) with high, medium, or low effort. Gemini 3 Flash Preview makes reasoning optional, with an extra minimal effort tier below low, both defaulting to medium.