Fireworks

    Llama 3.1 70b Model Card

    by Fireworks

    Explore detailed LLM evaluation metrics and benchmarks for Llama 3.1 70b, a open-source large language model by Fireworks. Compare API costs, context windows, and performance across multiple LLM benchmarks including MMLU (general knowledge), MMMU (multimodal capabilities), HellaSwag (common sense reasoning), HumanEval (code generation), and MATH (mathematical reasoning).

    Pricing

    Input Processing
    Cost per million tokens for processing prompts
    $0.90per million tokens
    Generation
    Cost per million tokens for responses
    $0.90per million tokens

    Model Details

    Context Length
    Maximum input length the model can process at once
    128000Ktokens
    Response Length
    Maximum length of a single generated response
    2Ktokens
    Release Date
    First public release date
    July 23, 2024
    Training Cutoff
    Latest date of data included in training
    December 2023
    License Type
    Model licensing and availability
    Open Source

    Performance Metrics

    General Knowledge (MMLU)
    Accuracy on diverse academic and real-world topics
    83.6score
    5-shot
    Multimodal Tasks (MMMU)
    Performance on tasks combining text, images, and other media
    Not available
    Common Sense (HellaSwag)
    Understanding of everyday scenarios and situations
    Not available
    Code Generation (HumanEval)
    Ability to write and understand programming code
    Not available
    Mathematical Ability (MATH)
    Accuracy in mathematical computations and reasoning
    68.0score
    0-shot
    AnotherWrapper Logo

    Building AI startups?

    You can speed up development time 10x using our 10+ Next.js AI templates.

    • 🚀

      10+ AI Templates

      Ready-to-use demos for text, image & chat

    • Modern Tech Stack

      Next.js 14, TypeScript & Tailwind

    • 🔌

      AI Integrations

      OpenAI, Anthropic & Replicate ready

    • 🛠️

      Full Infrastructure

      Auth, database & payments included

    • 🎨

      Professional Design

      6+ landing pages & modern UI kit

    • 📱

      Production Ready

      SEO optimized & ready to deploy

    All Large Language Models

    OpenAI

    GPT-3.5-turbo
    GPT-3.5-turbo-0125
    GPT-3.5-turbo-1106
    GPT-3.5-turbo-16k
    GPT-4 Turbo
    GPT-4-0613
    GPT-4-32k
    GPT-4-32k-0613
    GPT-4-turbo-0125
    GPT-4-turbo-1106
    GPT-4-turbo-2024-04-09
    GPT-4o
    GPT-4o-2024-05-13
    GPT-4o-2024-08-06
    GPT-4o-mini
    o1 mini
    o1 preview
    o1-mini-2024-09-12
    o1-preview-2024-09-12

    Anthropic

    Claude 2
    Claude 2.1
    Claude 3 Haiku
    Claude 3 Opus
    Claude 3 Sonnet
    Claude 3.5 Haiku
    Claude 3.5 Sonnet
    Claude Instant 1.2

    01.ai

    Yi-Large

    Anyscale

    Llama 3 70b
    Mixtral 8x7B

    Cohere

    Command R
    Command R+

    Deepseek

    deepseek-chat
    deepseek-coder

    Fireworks

    Llama 3.1 405b
    Mixtral 8x7B
    Yi-Large

    Google

    Gemini 1.5 Flash
    Gemini 1.5 Pro

    Groq

    Llama 3.1 70b
    Mixtral 8x7B

    IBM WatsonX

    Llama 3.1 405b
    Llama 3.1 70b
    Mistral Large

    Mistral

    Codestral
    Mistral 7B Instruct
    Mistral Large
    Mistral Large 2
    Mistral Nemo

    Perplexity

    Llama 3.1 70b

    Replicate

    Llama 3.1 405b
    Llama 3.1 70b
    Mixtral 8x7B

    Together.AI

    Llama 3.1 405b
    Llama 3.1 70b
    Mixtral 8x7B

    Writer

    Palmyra Fin 32k
    Palmyra Med 32k
    Palmyra X 002
    Palmyra X 002 32k
    Palmyra X 003
    Palmyra X 32k
    Next.js + OpenAI + Supabase + Replicate + Groq + GPT + LlaMA3 + TailwindCSS + Demo Apps

    Stop wasting time on infrastructure

    Focus on building a great product, not on setting up infrastructure. AnotherWrapper gives you the foundation to launch your AI product in days, not months.

    • 10+ ready-to-use AI demo apps
    • Pre-configured APIs and AI models
    • All essential SaaS features (auth, database, payments, etc.)

    FAQ
    Frequently asked questions

    Need help with something? Here are some of the most common questions we get.

    Still have questions? Email us at [email protected]