Llama 3.1 Tulu3 405B

Llama 3.1 Tulu3 405B

Allen Institute for AI · Released Jan 30, 2025 Emerging
Intelligence #505 / 583
15.2 Our Score
AA Index #260 / 380
8.3 Artificial Analysis
Input
— Not priced
Output
— Not priced
Context
— Not reported

Analysis Summary

Llama 3.1 Tulu3 405B is Allen Institute for AI's instruction-tuned variant of Meta's Llama 3.1 405B, released in early 2025. With an intelligence index of 8.3, MMLU-Pro of 0.716, and GPQA of 0.516, it sits in the upper-mid tier of the open-weight landscape and competes reasonably with commercial models from 2024.

For business use, its 405B scale makes it capable on complex reasoning, long-form content, and structured analysis tasks. Coding capability is moderate, with a LiveCodeBench score of 0.291, making it adequate for lighter engineering support but not a specialist coding model. The main constraint is infrastructure: running 405B parameters requires significant compute, which limits its accessibility for smaller teams.

Organisations with the infrastructure to self-host large models and a need for open-weight flexibility will find it a strong general-purpose option. For teams without that infrastructure, a hosted commercial model will deliver better cost-performance.

Assessed June 17, 2026

Editorial notes

Llama 3.1 Tulu3 405B from Allen Institute for AI is a large open-weight model with strong MMLU-Pro and GPQA scores, offering competitive general reasoning for self-hosted enterprise deployments.

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence2.1Technical0Value0Content5.5
Intelligence 2.1/10
Technical 0/10
Content 5.5/10
Value 0/10

How Llama 3.1 Tulu3 405B compares

Llama 3.1 Tulu3 405B ranks #260 of 380 AI models we track for overall intelligence. Llama 3.1 Tulu3 405B is currently free to use via OpenRouter.

Performance Indices

Source: Artificial Analysis

8.3 Intelligence Index

Benchmark Scores

Intelligence

GPQA Diamond 51.6% Graduate-level scientific reasoning
HLE 3.5% Humanity's Last Exam
MMLU Pro 71.6% Multi-task language understanding
MATH 500 77.8% Mathematical problem-solving
AIME 13.3% Competition mathematics
SciCode 30.2% Scientific computing

Technical

LiveCodeBench 29.1% Live coding evaluation

Benchmark data from Artificial Analysis and Hugging Face

How does Llama 3.1 Tulu3 405B stack up?

Compare side-by-side with other emerging models.

Compare Models

Model Information

ProviderAllen Institute for AI
Release Date January 30, 2025
Status Active

Leaderboard Categories

Frequently asked questions about Llama 3.1 Tulu3 405B

How much does Llama 3.1 Tulu3 405B cost?

Llama 3.1 Tulu3 405B is currently available for free via OpenRouter.

Who created Llama 3.1 Tulu3 405B?

Llama 3.1 Tulu3 405B is developed by Allen Institute For AI and was released on January 30, 2025.