AIbase
EN
Home
AI Models Finder
Jamba Instruct
J

Jamba Instruct

Early instruction-tuned version of AI21's Jamba using innovative SSM-Transformer hybrid architecture. Optimized for instruction following and conversation tasks, efficiently handles long context and complex instructions. Hybrid architecture provides unique efficiency advantages, reducing computational costs while maintaining high-quality output, suitable for enterprise conversation systems, document processing, and professional applications requiring long context understanding.
Intelligence(Relatively Weak)
Speed(Slow)
Input Supported Modalities
Yes
Is Reasoning Model
256,000
Context Window
-
Maximum Output Tokens
-
Knowledge Cutoff
Go Compare

Pricing

- /M tokens
Input
- /M tokens
Output
- /M tokens
Blended Price

Quick Simple Comparison

Input

Output

Jamba 1.7 Large
￥0.49
Jamba 1.7 Mini
Jamba 1.5 Mini
￥0.2

Basic Parameters

Jamba InstructTechnical Parameters
Parameter Count
Not Announced
Context Length
256.00k tokens
Training Data Cutoff
Open Source Category
Proprietary
Multimodal Support
Text Only
Throughput
Release Date
2024-05-02
Response Speed
0 tokens/s

Benchmark Scores

Below is the performance of Jamba Instruct in various standard benchmark tests. These tests evaluate the model's capabilities in different tasks and domains.
Intelligence Index
26.3
Large Language Model Intelligence Level
Coding Index
6.44
Indicator of AI model performance on coding tasks
Math Index
-
Capability indicator in solving mathematical problems, mathematical reasoning, or performing math-related tasks
MMLU Pro
34.3
Massive Multitask Multimodal Understanding - Testing understanding of text, images, audio, and video
GPQA
27.1
Graduate Physics Questions Assessment - Testing advanced physics knowledge with diamond science-level questions
HLE
5.2
The model's comprehensive average score on the Hugging Face Open LLM Leaderboard
LiveCodeBench
4.6
Specific evaluation focused on assessing large language models' ability in real-world code writing and solving programming competition problems
SciCode
8.3
The model's capability in code generation for scientific computing or specific scientific domains
HumanEval
-
Score achieved by the AI model on the specific HumanEval benchmark test set
Math 500 Score
24.3
Score on the first 500 larger, more well-known mathematical benchmark tests
AIME Score
-
An indicator measuring an AI model's ability to solve high-difficulty mathematical competition problems (specifically AIME level)
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase