Llama 3.1 Instruct 405B

Meta·Released 2024-07-23

Open Source

Compare

Quality Index

17.4

225th of 440

Top 51%

Coding Index

14.5

196th of 350

Top 57%

Math Index

3.0

256th of 268

Top 96%

Price/1M

$3.69

597th cheapest

1090% above median

Top 88%

Speed

37 tok/s

Top 55%

TTFT

0.56s

Market Position

Llama 3.1 Instruct 405BMarket Average

Pricing

Input

$2.75

per 1M tokens

Output

$6.50

per 1M tokens

Blended

$3.69

per 1M tokens

Cheaper than 12% of models. Median price is $0.31/1M tokens.

Cost Calculator

Tokens per day1M

100K100M

Daily

$3.69

Monthly

$110.64

vs. Similar Models

GPT-4o (Nov '24)Q:-0.1

$4.38+19%

Qwen3 VL 32B InstructQ:-0.2

$1.23-67%

DeepSeek R1 Distill Qwen 32BQ:-0.2

$0.27-93%

Gemini 2.5 Flash-Lite (Reasoning)Q:+0.2

$0.17-95%

Performance

37

tokens/sec

Faster than 45% of models

0.56

seconds

Faster than 42% of models

Market Median

46 tok/s

19% slower

Median TTFT

0.43s

29% slower

Throughput/Dollar

10

tok/s per $/1M

Speed Comparison

GPT-4

37 tok/s+1%

Claude 4 Opus (Non-reasoning)

37 tok/s-2%

Llama 3.1 Nemotron Instruct 70B

36 tok/s-3%

Benchmarks

MMLU-Pro

73.2%

GPQA Diamond

51.5%

HLE

4.2%

LiveCodeBench

30.5%

SciCode

29.9%

TerminalBench Hard

6.8%

MATH-500

70.3%

AIME

21.3%

AIME 2025

3.0%

IFBench

39.0%

Long Context Recall

24.3%

Tau2

19.0%

Market AverageTop Score

Open Source

Quick Compare

Similar Models

GPT-4o (Nov '24)

OpenAI

Q: 17.3$4.38/1M128K ctx

Faster: 235%Pricier: 19%

Qwen3 VL 32B Instruct

Qwen

Q: 17.2$1.23/1M131K ctx

Faster: 106%Cheaper: 67%

DeepSeek R1 Distill Qwen 32B

deepseek-ai

Q: 17.2$0.27/1M33K ctx

Faster: 64%Cheaper: 93%

Hermes 4 - Llama-3.1 405B (Non-reasoning)

Nous Research

Q: 17.6$1.50/1M

Slower: 21%Cheaper: 59%

Gemini 2.5 Flash-Lite (Reasoning)

Google

Q: 17.6$0.17/1M

Faster: 627%Cheaper: 95%

GLM-4.6V (Non-reasoning)

Z AI

Q: 17.1$0.45/1M131K ctx

Slower: 42%Cheaper: 88%

Compare all 7 models