AIbase
EN
Home
AI Models Finder
Jamba 1.7 Large
J

Jamba 1.7 Large

AI21's latest hybrid SSM-Transformer architecture model has 94B active parameters and 398B total parameters, and supports a 256K context window. It improves information baselining and instruction-following capabilities on the basis of basic performance, and uses ExpertsInt8 quantization technology, which can run efficiently on 8×80GB GPUs. It is suitable for enterprise-level long document analysis, RAG optimization, and complex dialogue tasks.
Intelligence(Relatively Weak)
Speed(Relatively Slow)
Input Supported Modalities
Yes
Is Reasoning Model
260,000
Context Window
0
Maximum Output Tokens
-
Knowledge Cutoff
Go Compare

Pricing

￥3.53 /M tokens
Input
￥7.99 /M tokens
Output
￥25.2 /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 1.7 LargeTechnical Parameters
Parameter Count
Not Announced
Context Length
260.00k tokens
Training Data Cutoff
Open Source Category
Open Weights (Permissive License)
Multimodal Support
Text Only
Throughput
0
Release Date
2025-07-07
Response Speed
59.83,873 tokens/s

Benchmark Scores

Below is the performance of Jamba 1.7 Large in various standard benchmark tests. These tests evaluate the model's capabilities in different tasks and domains.
Intelligence Index
30
Large Language Model Intelligence Level
Coding Index
18.42
Indicator of AI model performance on coding tasks
Math Index
32.83
Capability indicator in solving mathematical problems, mathematical reasoning, or performing math-related tasks
MMLU Pro
57.7
Massive Multitask Multimodal Understanding - Testing understanding of text, images, audio, and video
GPQA
39
Graduate Physics Questions Assessment - Testing advanced physics knowledge with diamond science-level questions
HLE
3.8
The model's comprehensive average score on the Hugging Face Open LLM Leaderboard
LiveCodeBench
18.1
Specific evaluation focused on assessing large language models' ability in real-world code writing and solving programming competition problems
SciCode
18.8
The model's capability in code generation for scientific computing or specific scientific domains
HumanEval
71
Score achieved by the AI model on the specific HumanEval benchmark test set
Math 500 Score
60
Score on the first 500 larger, more well-known mathematical benchmark tests
AIME Score
5.7
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