Gemini 3.5 Flash-Lite
Made by Google DeepMind, released 21 Jul 2026. Epoch AI lists its access as “API access”. No Top tier or Strong results yet among the models listed here.
In short
What it does well
No Top tier or Strong ratings on the four main tests yet.
Where it falls short
- Coding: Behind the leaders (68th of 77 listed models)
- Maths: Behind the leaders (42nd of 43 listed models)
- Writing: Middle (56th of 77 listed models)
What we don't know yet
- Short factual questions: Not enough data
- Fixing real software bugs: Not enough data
- Difficult coding tasks: Not enough data
The four ratings
- Top tier
- Strong
- Middle
- Behind the leaders
- Not enough data
Ratings compare only the 77 models listed here, in quarters, best first. “Not enough data” means no result yet. It is never a low score.
- Writing
- MiddleHow often people preferred its answers to writing prompts. 1,421 on Arena's scale · Source for Gemini 3.5 Flash-Lite writing score (lmarena)
- Coding
- Behind the leadersHow often people preferred its answers to coding questions. 1,453 on Arena's scale · Source for Gemini 3.5 Flash-Lite coding score (lmarena)
- Maths
- Behind the leadersHow many very hard maths problems it solved. 26.0% · Source for Gemini 3.5 Flash-Lite maths score (epoch)
- Knowledge
- Not enough dataNo result yet for short factual questions.
Evidence completeness: Broad evidence, 10 of 13 tests have a result.
Where it is relatively stronger or weaker
This compares each kind of request with the model's own overall position among the 77 listed models, using Arena's votes. It is about performance on different languages, kinds of request and topics, not about political or social bias.
Relatively stronger
- Japanese: 31st of 54 on questions asked in Japanese, against 65th of 77 overall. Source for Japanese
Relatively weaker
No category stands out as clearly weaker than its overall position.
Show every category we can see
| Category | Position | Arena score | Votes | Source |
|---|---|---|---|---|
| Expert-level prompts Kind of request | 71st of 77 | 1,434 95% range 1,424 to 1,444 | 4,183 | Source for Expert-level prompts |
| Hard prompts Kind of request | 68th of 77 | 1,441 95% range 1,436 to 1,446 | 23,729 | Source for Hard prompts |
| Long prompts Kind of request | 63rd of 77 | 1,435 95% range 1,429 to 1,441 | 17,043 | Source for Long prompts |
| Longer conversations Kind of request | 58th of 77 | 1,445 95% range 1,437 to 1,454 | 5,877 | Source for Longer conversations |
| Chinese Language | 67th of 76 | 1,467 95% range 1,455 to 1,479 | 2,728 | Source for Chinese |
| English prompts Language | 67th of 77 | 1,438 95% range 1,432 to 1,444 | 14,610 | Source for English prompts |
| French Language | 61st of 72 | 1,450 95% range 1,431 to 1,469 | 1,098 | Source for French |
| German Language | 37th of 55 | 1,450 95% range 1,427 to 1,473 | 660 | Source for German |
| Japanese Language | 31st of 54 | 1,428 95% range 1,402 to 1,454 | 594 | Source for Japanese |
| Korean Language | 43rd of 54 | 1,408 95% range 1,385 to 1,431 | 717 | Source for Korean |
| Non-English prompts Language | 63rd of 77 | 1,427 95% range 1,422 to 1,433 | 21,126 | Source for Non-English prompts |
| Russian Language | 55th of 77 | 1,445 95% range 1,435 to 1,455 | 4,041 | Source for Russian |
| Spanish Language | 54th of 70 | 1,442 95% range 1,422 to 1,461 | 958 | Source for Spanish |
| Entertainment, sport and media Topic | 56th of 77 | 1,415 95% range 1,408 to 1,422 | 9,216 | Source for Entertainment, sport and media |
| Law and government Topic | 66th of 76 | 1,431 95% range 1,419 to 1,442 | 3,038 | Source for Law and government |
| Medicine and health Topic | 60th of 75 | 1,439 95% range 1,427 to 1,451 | 2,597 | Source for Medicine and health |
| Science Topic | 61st of 77 | 1,451 95% range 1,443 to 1,460 | 5,844 | Source for Science |
| Software and IT Topic | 69th of 77 | 1,451 95% range 1,444 to 1,457 | 14,365 | Source for Software and IT |
Bias: what we can and can't see
Our sources do not measure political lean, stereotyping, flattery (agreeing with you too readily) or safety behaviour, so we cannot say how any model does on those. What we can show is above: where a model is relatively stronger or weaker across languages, kinds of request and topics. Also, people voting in Arena can favour answer style, such as length and formatting, and our numbers do not adjust for that. See what AI bias means and what these tests can tell you.
Best-supported use cases
- Research and checking facts: not ranked, because there is not enough data for the main test.
Where an alternative might be more suitable
- Writing and creating content: rated Middle on writing votes. Compare it with Claude Opus 5.5, the best-supported choice for this task.
- Coding and building things: rated Behind the leaders on coding votes. Compare it with Claude Opus 5.5, the best-supported choice for this task.
- Homework and maths: rated Behind the leaders on hard maths problems. Compare it with GPT-6.1 Sol, the best-supported choice for this task.
- Business emails and everyday work: rated Behind the leaders on business-task votes. Compare it with Gemini 4 Argon, the best-supported choice for this task.
- Research and checking facts: no result on short factual questions, so we cannot say. See which models the evidence does support.
What it is
- Maker
- Google DeepMind
- Release date (Epoch AI)
- 21 Jul 2026
- Access, as listed by Epoch AI
- API access
- Licence, as listed by Arena
- Proprietary
- Maker's country (Epoch AI)
- United States of America
- Last observation in our data
- 28 Oct 2026
- Data last changed
- 29 Oct 2026, 14:36 UTC
We have not verified an official product page for this model, so we do not link to one.
How you can use it
Proprietary. The model's files are not published, so it is generally available only through its maker or partners the maker chooses.
How you can reach it (an app, a website, a developer API or a third-party tool) and what that costs are not covered by our benchmark data. Where we could verify prices from the maker's own pages, they are shown under Price. Epoch AI lists access as “API access”. See models, apps and harnesses and how AI pricing works.
Price
Price not independently verified
See Gemini plans and pricing (checked 28 Oct 2026). Page opened and read.
We never infer subscription prices from API prices. New to “tokens” and “per million”? See how AI pricing works.
Good to remember
- Benchmarks are samples of tasks. None proves performance in every situation.
- Arena scores come from people choosing between answers. That measures preference, not correctness.
- Ratings are relative to the 77 models listed here. A different set would change them.
- Sources test models at different settings. Each result shows the setting used, and we do not adjust it.
The evidence, for the curious
Show all 13 tests with settings and sources
| Test | Result | Rating | Setting used | Source identifier | Observed | Links |
|---|---|---|---|---|---|---|
| Arena text: overall preference Arena 'text' leaderboard (Bradley-Terry rating, no style control) |
1,435 Arena score 95% range 1,431 to 1,440, 35,740 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: overall preference score (lmarena) |
| Arena text: creative writing Arena 'text' leaderboard, category creative_writing |
1,421 Arena score 95% range 1,413 to 1,429, 7,405 votes |
Middle | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: creative writing score (lmarena) |
| Arena text: coding Arena 'text' leaderboard, category coding |
1,453 Arena score 95% range 1,446 to 1,460, 9,922 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: coding score (lmarena) |
| Arena text: maths Arena 'text' leaderboard, category math |
1,436 Arena score 95% range 1,421 to 1,450, 1,626 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: maths score (lmarena) |
| Arena text: following instructions Arena 'text' leaderboard, category instruction_following |
1,420 Arena score 95% range 1,414 to 1,426, 12,867 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: following instructions score (lmarena) |
| Arena text: business and management Arena 'text' leaderboard, category industry_business_and_management_and_financial_operations |
1,423 Arena score 95% range 1,415 to 1,431, 6,600 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena text: business and management score (lmarena) |
| Arena WebDev: building web apps Arena 'webdev' leaderboard (Code Arena), category overall |
1,434 Arena score 95% range 1,393 to 1,475, 236 votes |
Behind the leaders | effort: not stated, bradley-terry rating | Gemini 3.5 Flash-Lite lmarena · webdev/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash-Lite Arena WebDev: building web apps score (lmarena) |
| FrontierMath Tiers 1-3 (v2) FrontierMath Tiers 1-3 v2 (private set), benchmark released 2026-06-12 |
26.0% share of problems solved standard error ±2.6 points |
Behind the leaders | effort: high, best score across scorers | gemini-3.5-flash-lite_high epoch · frontiermath_tiers_1_3_v2.csv |
2 Aug 2026 | Source for Gemini 3.5 Flash-Lite FrontierMath Tiers 1-3 (v2) score (epoch) · Run details for FrontierMath Tiers 1-3 (v2) |
| OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025, benchmark released 2024-12-19 |
71.1% share of problems solved standard error ±6.8 points |
Behind the leaders | effort: high, best score across scorers | gemini-3.5-flash-lite_high epoch · otis_mock_aime_2024_2025.csv |
6 Aug 2026 | Source for Gemini 3.5 Flash-Lite OTIS Mock AIME 2024-2025 score (epoch) · Run details for OTIS Mock AIME 2024-2025 |
| GPQA Diamond GPQA Diamond, benchmark released 2023-11-20 |
83.3% share of questions correct standard error ±2.7 points |
Behind the leaders | effort: high, best score across scorers | gemini-3.5-flash-lite_high epoch · gpqa_diamond.csv |
6 Aug 2026 | Source for Gemini 3.5 Flash-Lite GPQA Diamond score (epoch) · Run details for GPQA Diamond |
| SimpleQA Verified | Not enough data | |||||
| SWE-bench Verified (Epoch runs) | Not enough data | |||||
| FrontierCode (reported by Cognition) | Not enough data | |||||
Missing evidence
Not enough data for these tests:
- SimpleQA Verified
- SWE-bench Verified (Epoch runs)
- FrontierCode (reported by Cognition)
Missing results are not low scores and do not count against the model.
Where a source gives no public run log, we cannot show the run behind the score. See data sources.
Not sure this is the right one?
Tell us what you need help with and we'll narrow the options down.