The Model Index
Up to four models

GPT-4 (Jun 2023) vs Mixtral 8x22B

On the index (GPT-4 = 100), GPT-4 (Jun 2023) scores 102 and Mixtral 8x22B 99. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to GPT-4 (Jun 2023)
Presets
GPT-4 (Jun 2023)
10285–119
WeightsClosed
ReleasedJun 2023
Price in / out$30 / $60
Context8K
Core tests taken1 of 9
Mixtral 8x22B
9983–116
WeightsOpen
ReleasedApr 2024
Price in / out$2.0 / $6.0
Context66K
Core tests taken1 of 9

GPT-4 (Jun 2023) and Mixtral 8x22B are too close to call (3 points apart, within the uncertainty), ahead on 0 of 1 shared core tests, at 13× the price.

Test by testthird-party results

GPT-4 (Jun 2023)Mixtral 8x22B
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)MMLU (HELM runs)MMLU (HELM runs)MATHSMATH (level 5)MATH (level 5)CODINGWeirdMLWeirdMLAGENTSDTBenchDTBench
Every result as a table
TestGPT-4 (Jun 2023)Mixtral 8x22B
Reasoning & knowledge
GPQA Diamond (Epoch AI)30.6%34.1%
MMLU (HELM runs)82.4%77.8%
Maths
MATH (level 5)23.0%24.2%
Coding
WeirdML12.4%3.2%
Agents
DTBench62.7%55.1%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG