What are the verified coding and reasoning benchmarks for Google: Gemini 3.5 Flash Lite?

Google: Gemini 3.5 Flash Lite achieves SWE-bench Verified: 52% (Flash-Lite), LiveCodeBench: 45%. Verified sources: Artificial Analysis Intelligence Index v4.3.2 (Artificial Analysis), Gemini 3.5 Flash-Lite System Report (Google DeepMind).

What is the recommended Swarm multi-agent topology for Google: Gemini 3.5 Flash Lite?

Swarm recommends Google: Gemini 3.5 Flash Lite as Finder Subagent (Tier 3: Extreme-Throughput Reflex Subagent). Provides instant file discovery and test telemetry streaming. Best paired with: Gemini 3.8 Flash or Claude Opus 5.5.

How does Google: Gemini 3.5 Flash Lite compare to rival frontier coding models on SWE-bench and pricing?

Google: Gemini 3.5 Flash Lite achieves SWE-bench Verified: 52% at $0.075/$0.3 per 1M tokens. Direct peer comparison: Muse Glimmer (48% SWE-bench, $0.08/$0.32); Mistral Medium 3.5 (45% SWE-bench, $0.2/$0.8); DeepSeek V4.1 Flash (56.2% SWE-bench, $0.3/$0.9); Gemini 3.8 Flash (68.4% SWE-bench, $0.15/$0.6).

Does Swarm isolate Google: Gemini 3.5 Flash Lite code edits from other agents?

Yes. Swarm provisions isolated Git worktrees for every session and subagent, preventing file race conditions and keeping your main working tree clean until verification passes.

Can Google: Gemini 3.5 Flash Lite be assigned to specific subagent roles?

Yes. In ~/.swarm/config.json, you can configure Google: Gemini 3.5 Flash Lite as Primary/Planner (DAG task decomposition), Coder (implementation in isolated worktrees), or Finder (read-only symbol search).

Google: Gemini 3.5 Flash Lite — Specs, Context & Capabilities

Q: How does Swarm isolate Google: Gemini 3.5 Flash Lite during autonomous code generation?

Scans worktrees in parallel without adding wall-clock latency.

Q: What are developers and practitioners saying about Google: Gemini 3.5 Flash Lite in production?

The absolute king of interactive latency and high-speed streaming. Developers use it for continuous real-time linting, formatting, and AST sanity checks.

Q: How do I run Google: Gemini 3.5 Flash Lite with Swarm?

Configure Google: Gemini 3.5 Flash Lite via CLI using: swarm ctl model set --provider openrouter --model openrouter/google/gemini-3.5-flash-lite, or switch models interactively inside the terminal TUI with /models.

Google: Gemini 3.5 Flash Lite Benchmarks & Agentic Accuracy

Verified primary research & developer preference metrics

Primary Sources Verified

SWE-bench Verified

52%

Flash-LiteGoogle DeepMind ↗

LiveCodeBench

45%

StandardLiveCodeBench ↗

LMSYS Coding Elo

1295Elo

Coding ArenaLMSYS Arena ↗

Pricing ($ / 1M tokens)

$0.075/ $0.3

$0.015 cached readStandard API Pricing

What This Means in Real Software Development

Outputs 320 tokens per second, ranking among the fastest models measured globally.
Achieves the lowest cost in Google's lineup at $0.05 per task ($0.075/M input).
Scores 22 on the Artificial Analysis Intelligence Index v4.3.2 and 24 on Finance & Accounting.
Near-instantaneous time to first token (<0.35s) for reflexive subagent tool checks.

Swarm Multi-Agent Topology

Tier 3: Extreme-Throughput Reflex SubagentRole: Finder Subagent

Best Paired With: Gemini 3.8 Flash or Claude Opus 5.5

Provides instant file discovery and test telemetry streaming.

Git Worktree Isolation Advantage:

Scans worktrees in parallel without adding wall-clock latency.

What Developers & Practitioners Are Saying

Real-world field observations, production strengths, and developer consensus

Field Consensus

"The absolute king of interactive latency and high-speed streaming. Developers use it for continuous real-time linting, formatting, and AST sanity checks."

Verified Real-World Strengths

Unmatched 320 TPS generation throughput with sub-second TTFT
Rock-bottom $0.05 cost per task and $0.075/M input price
Massive context capacity with prompt caching efficiency

Production Caveats & Gaps

Lower complex reasoning ceiling than Gemini 3.8 Flash
Not designed for autonomous multi-file architecture refactoring

Primary Sources & Methodology Citations

Artificial Analysis Intelligence Index v4.3.2 — Artificial Analysis (September 2026)
Gemini 3.5 Flash-Lite System Report — Google DeepMind (May 2026)

Head-to-Head Frontier Comparisons: Google: Gemini 3.5 Flash Lite vs Rivals

Standardized, side-by-side benchmark & economics comparison against peer models

Flagship Comparative Matrix

Model	Role / Tier	SWE-bench	LiveCodeBench	LMSYS Elo	AIME	Input / Output
Google: Gemini 3.5 Flash LiteOpenRouterCurrent	Tier 3: Extreme-Throughput Reflex Subagent	52%	45%	1295	—	$0.075 / $0.3
Muse GlimmerMeta Compare Google: Gemini 3.5 Flash Lite vs Muse Glimmer →	Tier 3: Rapid Transpiler Subagent	48%	41.5%	1270	—	$0.08 / $0.32
Mistral Medium 3.5Mistral Compare Google: Gemini 3.5 Flash Lite vs Mistral Medium 3.5 →	Tier 3: European Sovereign Subagent	45%	38%	1250	—	$0.2 / $0.8
DeepSeek V4.1 FlashDeepSeek Compare Google: Gemini 3.5 Flash Lite vs DeepSeek V4.1 Flash →	Tier 3: Refined Open Subagent	56.2%	52%	1360	74.5%	$0.3 / $0.9
Gemini 3.8 FlashGoogle Compare Google: Gemini 3.5 Flash Lite vs Gemini 3.8 Flash →	Tier 3: High-Throughput Subagent Engine	68.4%	53.8%	1380	84%	$0.15 / $0.6

Explore All AI Model Comparisons Hub (Google: Gemini 3.5 Flash Lite & Flagship Peers) →

Google: Gemini 3.5 Flash Lite model overview

Google: Gemini 3.5 Flash Lite is listed in the public Swarm catalog through OpenRouter. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.

Provider: OpenRouter
Family: OpenRouter
Context window: 1.05M tokens
Thinking support: low · medium · high
Availability status: Available
Catalog ID: openrouter/google/gemini-3.5-flash-lite

Audio input
File input
Image input
Text input
Video input
Text output
Reasoning
Tools
Structured output
JSON

Run Google: Gemini 3.5 Flash Lite in Swarm

Verified CLI workflow & multi-agent architecture

Quickstart Guide →

bash — swarm run

$swarm ctl model set --provider openrouter --model openrouter/google/gemini-3.5-flash-lite

# Or switch interactively inside the terminal TUI:

>/models

Subagent Role Matrix

Assign Google: Gemini 3.5 Flash Lite to specialized system roles in ~/.swarm/config.json:

Primary / Planner: Orchestrates full task DAGs and multi-stage pipelines.
Coder Subagent: Generates code and unit tests inside private Git worktrees.
Finder Subagent: Performs read-only code search, symbol discovery, and mapping.

System Agent Roles →Why Swarm Architecture →