Model

Gemini 2.5 Flash-Lite (Thinking)

CLOSED

Google · Gemini 2.5 family · released Jun 17, 2025

Same Flash-Lite weights with thinking budget enabled for stronger reasoning at extra cost.

ReasoningCodingVisionFunction callingTool useAgentic

1281.7

Elo · rank #161

Parameters

Undisclosed

Active params

Undisclosed

Context

1M tokens

Architecture

Sparse Mixture-of-Experts Transformer (natively multimodal)

License

Proprietary (Google API Terms)

Languages

—

API price (in/out)

$0.1 / $0.4

Modalities

text · vision · audio · video

Benchmark results

Bar shows position within the tracked field; marker = field best

Aider PolyglotCoding27.1%#32

best: Claude Opus 4.5 · 89.4%

AIMEMath63.1%#110

best: GPT-5.2 · 100.0%

FACTS GroundingKnowledge86.8%#1

best: this model · 86.8%

GPQA DiamondReasoning66.7%#134

best: GPT-5.6 · 94.6%

Humanity's Last ExamReasoning6.9%#106

best: Claude Sonnet 5 · 57.4%

LiveCodeBenchCoding34.3%#142

best: DeepSeek-V4-Pro (Think Max) · 93.5%

MMLUKnowledge84.5%#62

best: OpenAI o3 · 92.9%

MMMUVision72.9%#44

best: Claude Fable 5 · 89.3%

SimpleQAKnowledge13.0%#30

best: GPT-4.5 · 62.5%

SWE-bench VerifiedCoding27.6%#111

best: Claude Fable 5 · 95.0%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$0.1

Output / M tok

$0.4

Gemini 2.5 family

Elo progression across releases

Gemini 2.5 ProMar 20251865.2 Gemini 2.5 FlashMay 20251637.3 Gemini 2.5 Flash-Lite (Non-thinking)Jun 20251193.2 Gemini 2.5 Flash-Lite (Thinking)Jun 20251281.7

Resources

DOCSai.google.dev/gemini-api/docs/models/gemini-2.5-flash-lite APIAvailable via Google API

API price $0.1/$0.4 · each benchmark row carries its own source badge (see methodology)