Gemini 2.0 Flash Thinking: Complete Benchmark Performance Guide

google 2.0-flash-thinking 2024-12-19 Proprietary

In-depth analysis of Gemini 2.0 Flash Thinking performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendorgoogle
Version2.0-flash-thinking
Release Date2024-12-19
Context Window1.048576e+06 tokens
Input Modalitiestext, image
Output Modalitiestext
LicenseProprietary
Documentationhttps://ai.google.dev/gemini-api/docs

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU86.5%2024-12-195-shotview
HUMANEVAL87.2pass@12024-12-19—view
GSM8K91.3%2024-12-190-shot CoTview
MATH56.6%2024-12-190-shot CoTview
BBH87.9%2024-12-193-shot CoTview
GPQA61%2024-12-190-shotview
IFEVAL82.8%2024-12-19prompt_strictview
ARC95%2024-12-19challengeview
MUSR70.8%2024-12-190-shotview
WINOGRANDE88.1%2024-12-190-shotview

Pricing

TierPriceCurrency
Input$0.1 / MtokUSD
Output$0.4 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://ai.google.dev/pricing · as of 2024-12-19

Compliance

  • Data Residency: US
  • SOC2: ✓
  • HIPAA: ✗
  • GDPR: ✓
  • ISO 27001: ✓

Gemini 2.0 Flash Thinking

모델 개요

Google Gemini 2.0 Flash Thinking 实验版, 1M 上下文, 链式思维推理, 在数学与编码上接近 o1。

핵심 사양

공급업체버전출시일컨텍스트 창입력 모달리티출력 모달리티라이선스
Google2.0-flash-thinking2024-12-191048Ktext, imagetextProprietary

벤치마크 성능

벤치마크점수단위비고
MMLU (Massive Multitask Language Understanding)86.5%5-shot
HumanEval87.2pass@1—
GSM8K (Grade School Math 8K)91.3%0-shot CoT
MATH56.6%0-shot CoT
BBH (BIG-Bench Hard)87.9%3-shot CoT
GPQA61.0%0-shot
IFEval82.8%prompt_strict
ARC95.0%challenge
MUSR70.8%0-shot
WinoGrande88.1%0-shot

가격

입력출력캐시 읽기캐시 쓰기
————

백만 토큰당

강점

  • MMLU score 86.5, strong knowledge reasoning.
  • HumanEval 87.2, excellent code generation.
  • GSM8K 91.3, robust math reasoning.
  • 支持文本、图像、音频多模态输入。

약점

  • 闭源专有模型，不支持自托管。

사용 사례

  • 代码生成与调试
  • 视觉与图像理解
  • Agent 工作流与工具调用

참고문헌