Key Specifications

Vendorgoogle
Version1.5-flash-8b-002
Release Date2024-10-03
Context Window1e+06 tokens
Input Modalitiestext, image, audio, video
Output Modalitiestext
LicenseProprietary
Documentationhttps://ai.google.dev/gemini-api/docs

Benchmark Performance

Benchmark Score Unit Evaluated At Notes Source
MMLU 77.5 % 2024-10-03 5-shot view
HUMANEVAL 79.7 pass@1 2024-10-03 view
GSM8K 86.3 % 2024-10-03 0-shot CoT view
MATH 43 % 2024-10-03 0-shot CoT view
BBH 73.6 % 2024-10-03 3-shot CoT view
GPQA 39.8 % 2024-10-03 0-shot view
IFEVAL 74.3 % 2024-10-03 prompt_strict view
ARC 93.2 % 2024-10-03 challenge view
MUSR 54.7 % 2024-10-03 0-shot view
WINOGRANDE 84.6 % 2024-10-03 0-shot view

Pricing

Tier Price Currency
Input$0.0375 / MtokUSD
Output$0.15 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://ai.google.dev/pricing · as of 2024-10-03

Compliance

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

Gemini 1.5 Flash-8B 002

Model Overview

Google Gemini 1.5 Flash-8B 002 生产版本, 1M 上下文, 最低成本多模态推理。

Core Specifications

Vendor Version Release Date Context Window Input Modalities Output Modalities License
Google 1.5-flash-8b-002 2024-10-03 1000K text, image, audio, video text Proprietary

Benchmark Performance

Benchmark Score Unit Notes
MMLU (Massive Multitask Language Understanding) 77.5 % 5-shot
HumanEval 79.7 pass@1
GSM8K (Grade School Math 8K) 86.3 % 0-shot CoT
MATH 43.0 % 0-shot CoT
BBH (BIG-Bench Hard) 73.6 % 3-shot CoT
GPQA 39.8 % 0-shot
IFEval 74.3 % prompt_strict
ARC 93.2 % challenge
MUSR 54.7 % 0-shot
WinoGrande 84.6 % 0-shot

Pricing

Input Output Cache Read Cache Write

per million tokens

Strengths

  • GSM8K 86.3, robust math reasoning.
  • Input Modalities: text, image, audio.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging
  • Vision and image understanding

References