Jamba 1.5: Complete Benchmark Performance Guide

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

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

Key Specifications

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

Benchmark Performance

BenchmarkScoreUnitEvaluated AtNotesSource
MMLU78.5%2024-08-075-shotview
HUMANEVAL75.9pass@12024-08-07—view
GSM8K79.8%2024-08-070-shot CoTview
MATH48.1%2024-08-070-shot CoTview
BBH77.3%2024-08-073-shot CoTview
GPQA30.2%2024-08-070-shotview
IFEVAL70.8%2024-08-07prompt_strictview
ARC94.2%2024-08-07challengeview
MUSR57.1%2024-08-070-shotview
WINOGRANDE84.5%2024-08-070-shotview

Pricing

TierPriceCurrency
Input$0.5 / MtokUSD
Output$0.7 / 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

Model Overview

AI21 Labs Jamba 1.5 基础模型系列, 256K 上下文, 混合 SSM-Transformer 架构, 高效长上下文处理。

Core Specifications

VendorVersionRelease DateContext WindowInput ModalitiesOutput ModalitiesLicense
Otherjamba-1-52024-08-07256KtexttextJamba Open Model License

Benchmark Performance

BenchmarkScoreUnitNotes
MMLU (Massive Multitask Language Understanding)78.5%5-shot
HumanEval75.9pass@1—
GSM8K (Grade School Math 8K)79.8%0-shot CoT
MATH48.1%0-shot CoT
BBH (BIG-Bench Hard)77.3%3-shot CoT
GPQA30.2%0-shot
IFEval70.8%prompt_strict
ARC94.2%challenge
MUSR57.1%0-shot
WinoGrande84.5%0-shot

Pricing

InputOutputCache ReadCache Write
————

per million tokens

Strengths

  • Context window 256K.

Weaknesses

  • Proprietary, not self-hostable.

Use Cases

  • Code generation and debugging
  • Long document summarization

References