Gemini 3.5 Flash: benchmarks, price and rank
Gemini 3.5 Flash from Google is 47th of 123 rated AI models on Spezia's leaderboard, strongest on MMMU-Pro and Finance Agent v2. It costs $1.50 in, $9 out per million tokens.
- 47th of 123 rated models
- Rating 1,367
- Rated on 18 benchmarks
- Closed
- Released May 19, 2026
- 1 million tokens of context
- Knows the world to January 2026
Where it is strongest
- 2nd of 51 on MMMU-Pro
- 1st of 23 on Finance Agent v2
Where it falls behind
- 52nd of 56 on APEX-Agents
- 17th of 24 on MRCR v2 (8 needles)
- 35th of 50 on SWE-Bench Pro
Price: $1.50 in, $9 out per million tokens.
Its headline results
The benchmarks that count most in the rating, with its place among the models measured the same way, and where each number came from.
- APEX-Agents27.5%
Long tasks written by investment banking analysts, management consultants and corporate lawyers from their own work, done across files and apps in a realistic work setup.
52nd of 56. Independent result from Epoch AI.
- MCP Atlas83.6%
Multi-step tasks where the model has to pick and combine many real tools, connected through MCP, the standard way agents plug into apps and data.
5th of 33. The maker's own number, from Google, via LLM Stats.
- SWE-Bench Pro55.1%
Long changes to real codebases in several languages, often across many files, including code from private and copyleft projects that models are unlikely to have seen.
35th of 50. The maker's own number, from Google, via LLM Stats.
- Humanity's Last Exam40.2%
About 2,500 questions written by experts in more than a hundred fields to be as hard as they could make them, many needing a diagram or image.
29th of 74. The maker's own number, from Google, via LLM Stats.
- Vending-Bench 2$5,396
The model runs a simulated vending machine business for a year on its own: finding suppliers, ordering stock, setting prices and paying the daily fees.
31st of 60. Independent result from Epoch AI.
- OSWorld-Verified78.4%
Ordinary tasks on a real desktop computer using normal apps: a browser, a spreadsheet, an email client, a photo editor, the file manager.
10th of 25. The maker's own number, from Google, via LLM Stats.
- Toolathlon56.5%
Long tasks across many real apps and their tools, such as email, calendars, spreadsheets, code repositories and online stores.
11th of 37. The maker's own number, from Google, via LLM Stats.
- DeepSWE 1.137%
Software engineering tasks run in the same simple agent setup for every model.
Independent result from LLM Stats.
- SimpleBench76.7%
Multiple-choice puzzles about everyday situations, space, time and people, written to trip up models that pattern-match instead of think. Ordinary adults do well on them.
14th of 68. Independent result from Epoch AI.
- SimpleQA Verified66.2%
Short factual questions with one checked answer, asked without web search.
14th of 78. Independent result from Epoch AI.
13 more results in its full profile.
Every result with its source, what it costs per task, what jumped with its release, side by side with any other model, and answers to your questions about when to use it.
Models ranked near it
How the rating works, and every benchmark it uses: the model leaderboard.