---
title: "Gemini 3.5 Flash Benchmarks - RankedAGI"
description: "Detailed benchmark and metadata record for Gemini 3.5 Flash."
source: "https://rankedagi.com/models/gemini-3-5-flash"
---

# Gemini 3.5 Flash

| Field | Value |
| --- | --- |
| Organization | Google |
| License | Proprietary |
| Type | Thinking |
| Version | Latest |
| Released | 2026-05-19 |
| Context window | 1M |
| Knowledge cutoff | 2025-01-01 |
| Input cost per million tokens | $1.5 |
| Output cost per million tokens | $10 |
| Cached input cost per million tokens | $0.15 |
| Model ID | gemini-3.5-flash |
| Last updated | 2026-05-20T12:13:26.004Z |

## Benchmarks

| Benchmark | Category | Value | Description | Source |
| --- | --- | --- | --- | --- |
| RankedAGI Coding | coding | 85.6% | RankedAGI Coding Score |  |
| SWEBench Pro | coding | 55.1% | Diverse Agentic Coding |  |
| RankedAGI Agentic | Agents | 86.0% | RankedAGI Agentic Score |  |
| Code Arena | coding | 1507 | Code Arena Elo Score | https://arena.ai/leaderboard/code |
| OSWorld-Verified | Agents | 78.4% |  |  |
| RankedAGI Reasoning | reasoning | 72.9% | RankedAGI Reasoning Score |  |
| Humanity's Last Exam | reasoning | 40.2% | Multidisciplinary Reasoning (no tools) |  |
| MMMU-Pro | imaging | 83.6% | Multimodal Understanding |  |
| RankedAGI Math | math | 70.5% | RankedAGI Math Score |  |
| RankedAGI Overall | general | 71.6% | Overall RankedAGI score |  |
| ARC AGI 2 | reasoning | 72.1% | Abstract Reasoning Puzzles (Public) |  |
| GDPval-AA Elo | Agents | 1656 | Office Tasks (Artificial Analysis) |  |
| MCP Atlas | Agents | 83.6% | Scaled Tool Use |  |
| Finance Agent | Agents | 57.9% | Agentic Financial Analysis |  |
| CharXiv Reasoning | Multimodal | 84.2% | Figure Understanding |  |
| Toolathlon | Agents | 56.5% | Benchmark to assess language agents' general tool use in realistic environments. | https://toolathlon.xyz/introduction |
| Cursor Bench 3.1 | coding | 49.8% | Benchmark to evaluate agents on ambiguous, multi-file tasks from real Cursor sessions. | https://cursor.com/evals |

## Model Links

| Title | URL |
| --- | --- |
| DeepMind | https://deepmind.google/models/gemini/flash/ |
| Google | https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/ |
| Model Card | https://deepmind.google/models/model-cards/gemini-3-5-flash/ |