Jamba 1.5 Mini: Complete Benchmark Performance Guide

other jamba-1-5-mini 2024-08-07 Jamba Open Model License

In-depth analysis of Jamba 1.5 Mini performance on MMLU, HumanEval, GSM8K, MATH, BBH benchmarks, including pricing and use cases.

Key Specifications

Vendorother
Versionjamba-1-5-mini
Release Date2024-08-07
Context Window256000 tokens
Input Modalitiestext
Output Modalitiestext
LicenseJamba Open Model License
Documentationhttps://huggingface.co/models

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU77.9%2024-08-075-shotview
HUMANEVAL67pass@12024-08-07—view
GSM8K75.7%2024-08-070-shot CoTview
MATH42.5%2024-08-070-shot CoTview
BBH75.6%2024-08-073-shot CoTview
GPQA34.2%2024-08-070-shotview
IFEVAL71.9%2024-08-07prompt_strictview
ARC91.8%2024-08-07challengeview
MUSR48.1%2024-08-070-shotview
WINOGRANDE81%2024-08-070-shotview

Pricing

TierPriceCurrency
Input$0.3 / MtokUSD
Output$0.5 / MtokUSD
Cache Read$0 / MtokUSD
Cache Write$0 / MtokUSD

Source: https://huggingface.co/models · as of 2024-08-07

Compliance

  • Data Residency: self-host
  • SOC2: ✗
  • HIPAA: ✗
  • GDPR: ✗
  • ISO 27001: ✗

Jamba 1.5 Mini

Model Overview

AI21 Labs Jamba 1.5 Mini 混合模型, 256K 上下文, 总参 52B/活跃 12B, 经济型长上下文模型。

Core Specifications

VendorVersionRelease DateContext WindowInput ModalitiesOutput ModalitiesLicense
Otherjamba-1-5-mini2024-08-07256KtexttextJamba Open Model License

Benchmark Performance

BenchmarkScoreUnitNotes
MMLU (Massive Multitask Language Understanding)77.9%5-shot
HumanEval67.0pass@1—
GSM8K (Grade School Math 8K)75.7%0-shot CoT
MATH42.5%0-shot CoT
BBH (BIG-Bench Hard)75.6%3-shot CoT
GPQA34.2%0-shot
IFEval71.9%prompt_strict
ARC91.8%challenge
MUSR48.1%0-shot
WinoGrande81.0%0-shot

Pricing

InputOutputCache ReadCache Write
————

per million tokens

Strengths

  • Context window 256K.
  • Mixture-of-Experts architecture.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Long document summarization

References