Gemini 3.5 Flash
Made by Google DeepMind, released 19 May 2026. Epoch AI lists its access as “API access”. Appears strongest at knowledge and writing among the models listed here.
In short
What it does well
- Knowledge: Top tier (10th of 43 listed models)
- Writing: Top tier (15th of 77 listed models)
- Coding: Strong (29th of 77 listed models)
Where it falls short
- Maths: Middle (28th of 43 listed models)
What we don't know yet
- 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
- Top tierHow often people preferred its answers to writing prompts. 1,469 on Arena's scale · Source for Gemini 3.5 Flash writing score (lmarena)
- Coding
- StrongHow often people preferred its answers to coding questions. 1,493 on Arena's scale · Source for Gemini 3.5 Flash coding score (lmarena)
- Maths
- MiddleHow many very hard maths problems it solved. 62.8% · Source for Gemini 3.5 Flash maths score (epoch)
- Knowledge
- Top tierHow many short factual questions it answered correctly. 66.2% · Source for Gemini 3.5 Flash knowledge score (epoch)
Evidence completeness: Broad evidence, 12 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
No category stands out as clearly stronger than its overall position.
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 | 29th of 77 | 1,494 95% range 1,486 to 1,503 | 5,567 | Source for Expert-level prompts |
| Hard prompts Kind of request | 20th of 77 | 1,488 95% range 1,484 to 1,493 | 32,328 | Source for Hard prompts |
| Long prompts Kind of request | 20th of 77 | 1,483 95% range 1,478 to 1,488 | 22,730 | Source for Long prompts |
| Longer conversations Kind of request | 20th of 77 | 1,481 95% range 1,474 to 1,488 | 8,083 | Source for Longer conversations |
| Chinese Language | 22nd of 76 | 1,525 95% range 1,515 to 1,536 | 3,583 | Source for Chinese |
| English prompts Language | 19th of 77 | 1,481 95% range 1,475 to 1,486 | 20,531 | Source for English prompts |
| French Language | 21st of 72 | 1,489 95% range 1,473 to 1,506 | 1,546 | Source for French |
| German Language | 10th of 55 | 1,493 95% range 1,471 to 1,515 | 773 | Source for German |
| Japanese Language | 11th of 54 | 1,490 95% range 1,466 to 1,514 | 707 | Source for Japanese |
| Korean Language | 16th of 54 | 1,445 95% range 1,425 to 1,465 | 970 | Source for Korean |
| Non-English prompts Language | 13th of 77 | 1,476 95% range 1,471 to 1,480 | 27,843 | Source for Non-English prompts |
| Russian Language | 12th of 77 | 1,494 95% range 1,486 to 1,503 | 5,437 | Source for Russian |
| Spanish Language | 13th of 70 | 1,481 95% range 1,465 to 1,498 | 1,425 | Source for Spanish |
| Entertainment, sport and media Topic | 16th of 77 | 1,459 95% range 1,452 to 1,465 | 11,832 | Source for Entertainment, sport and media |
| Law and government Topic | 18th of 76 | 1,487 95% range 1,477 to 1,497 | 4,050 | Source for Law and government |
| Medicine and health Topic | 19th of 75 | 1,484 95% range 1,473 to 1,494 | 3,543 | Source for Medicine and health |
| Science Topic | 19th of 77 | 1,494 95% range 1,487 to 1,501 | 7,983 | Source for Science |
| Software and IT Topic | 23rd of 77 | 1,494 95% range 1,489 to 1,499 | 19,402 | 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
This model is not among the options we recommend for any of the five tasks. That reflects the rules and the models listed, not a verdict on the model.
Where an alternative might be more suitable
- Coding and building things: rated Strong on coding votes. Compare it with Claude Opus 5.5, the best-supported choice for this task.
- Homework and maths: rated Middle on hard maths problems. Compare it with GPT-6.1 Sol, the best-supported choice for this task.
What it is
- Maker
- Google DeepMind
- Release date (Epoch AI)
- 19 May 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) |
1,482 Arena score 95% range 1,478 to 1,486, 48,376 votes |
Top tier | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: overall preference score (lmarena) |
| Arena text: creative writing Arena 'text' leaderboard, category creative_writing |
1,469 Arena score 95% range 1,461 to 1,476, 9,379 votes |
Top tier | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: creative writing score (lmarena) |
| Arena text: coding Arena 'text' leaderboard, category coding |
1,493 Arena score 95% range 1,487 to 1,499, 13,850 votes |
Strong | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: coding score (lmarena) |
| Arena text: maths Arena 'text' leaderboard, category math |
1,504 Arena score 95% range 1,492 to 1,517, 2,385 votes |
Top tier | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: maths score (lmarena) |
| Arena text: following instructions Arena 'text' leaderboard, category instruction_following |
1,467 Arena score 95% range 1,462 to 1,473, 17,342 votes |
Strong | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: following instructions score (lmarena) |
| Arena text: business and management Arena 'text' leaderboard, category industry_business_and_management_and_financial_operations |
1,467 Arena score 95% range 1,460 to 1,474, 9,384 votes |
Top tier | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · text/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena text: business and management score (lmarena) |
| Arena WebDev: building web apps Arena 'webdev' leaderboard (Code Arena), category overall |
1,499 Arena score 95% range 1,492 to 1,507, 9,331 votes |
Middle | effort: High, bradley-terry rating | Gemini 3.5 Flash (High) lmarena · webdev/latest |
28 Oct 2026 | Source for Gemini 3.5 Flash Arena WebDev: building web apps score (lmarena) |
| FrontierMath Tiers 1-3 (v2) FrontierMath Tiers 1-3 v2 (private set), benchmark released 2026-06-12 |
62.8% share of problems solved standard error ±2.9 points |
Middle | effort: high, best score across scorers | gemini-3.5-flash_high epoch · frontiermath_tiers_1_3_v2.csv |
10 Jun 2026 | Source for Gemini 3.5 Flash 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 |
95.6% share of problems solved standard error ±2.7 points |
Middle | effort: high, best score across scorers | gemini-3.5-flash_high epoch · otis_mock_aime_2024_2025.csv |
25 May 2026 | Source for Gemini 3.5 Flash OTIS Mock AIME 2024-2025 score (epoch) · Run details for OTIS Mock AIME 2024-2025 |
| GPQA Diamond GPQA Diamond, benchmark released 2023-11-20 |
92.8% share of questions correct standard error ±1.6 points |
Strong | effort: high, best score across scorers | gemini-3.5-flash_high epoch · gpqa_diamond.csv |
22 May 2026 | Source for Gemini 3.5 Flash GPQA Diamond score (epoch) · Run details for GPQA Diamond |
| SimpleQA Verified SimpleQA Verified, benchmark released 2025-09-09 |
66.2% share of questions correct standard error ±1.5 points |
Top tier | effort: high, best score across scorers | gemini-3.5-flash_high epoch · simpleqa_verified.csv |
27 Aug 2026 | Source for Gemini 3.5 Flash SimpleQA Verified score (epoch) · Run details for SimpleQA Verified |
| SWE-bench Verified (Epoch runs) SWE-bench Verified, benchmark released 2024-08-13, as run by Epoch AI |
79.3% share of issues resolved standard error ±1.8 points |
Top tier | effort: high, best score across scorers | gemini-3.5-flash_high epoch · swe_bench_verified.csv |
1 Jun 2026 | Source for Gemini 3.5 Flash SWE-bench Verified (Epoch runs) score (epoch) · Run details for SWE-bench Verified (Epoch runs) |
| FrontierCode (reported by Cognition) | Not enough data | |||||
Missing evidence
Not enough data for these tests:
- 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.