Noometry

Model comparison

GPT-4 vs Mistral 7B

GPT-4 is the stronger model overall, scoring 29.1 to 23.0 on the Noometry Index. Mistral 7B costs 150× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.

Last verified . 31 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

Mistral 7B Mistral AI

23.0

Rank #351 Confirmed

Summary

  • They share 31 benchmarks with published results for both. GPT-4 scores higher in 8 categories and Mistral 7B in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in multilingual, where GPT-4 leads 40.6 to 25.8.
  • The biggest single-benchmark swing is BigCodeBench Complete: 57.2% for GPT-4 and 27.3% for Mistral 7B.
  • Mistral 7B is cheaper at $0.25 / $0.25 per million input/output tokens, against $30 / $60 for GPT-4.
  • GPT-4 accepts more context: 8K tokens versus 8K.
  • Mistral 7B has downloadable open weights; the other is API-only.

Side by side

GPT-4Mistral 7B
ProviderOpenAIMistral AI
Noometry Index29.123.0
Released2023-03-142023-09-27
WeightsProprietaryOpen
Context window8K8K
Max output8K8K
Input $ / M tokens$30$0.25
Output $ / M tokens$60$0.25
Results tracked3837

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding GPT-4 leads

GPT-4: 31.6 (#283), Mistral 7B: 26.4 (#326)

BenchmarkGPT-4Mistral 7B
BigCodeBench Instruct46%19.5%
LMArena Coding12541082
BigCodeBench Complete57.2%27.3%
HumanEval+79.3%36%
WeirdML12.4%—
MBPP+—42.1%

Agentic & Tool Use Not comparable

GPT-4: —, Mistral 7B: —

BenchmarkGPT-4Mistral 7B
METR Time Horizons36.1%—

Reasoning GPT-4 leads

GPT-4: 17.8 (#289), Mistral 7B: 13.1 (#336)

BenchmarkGPT-4Mistral 7B
Chess Puzzles4%0%
LMArena Hard Prompts12411067
DTBench62.7%42.5%
BIG-Bench Hard75.1%56.1%
Epoch Capabilities Index125.89112.21
HellaSwag95.3%81%
WinoGrande87.5%75.3%
Mystery Game Puzzles12%—
LMCA17.1%—
Adversarial NLI—47.1%
ForecastBench57.8—
PIQA—83%

Math GPT-4 leads

GPT-4: 10.8 (#309), Mistral 7B: 8.1 (#325)

BenchmarkGPT-4Mistral 7B
OTIS Mock AIME 2024-20251.1%0.3%
LMArena Math12691085
MATH Level 523%3.7%
GSM8K92%54.4%

Knowledge GPT-4 leads

GPT-4: 18.4 (#282), Mistral 7B: 7.4 (#311)

BenchmarkGPT-4Mistral 7B
GPQA Diamond35.7%15.2%
LMArena Expert12111036
MMLU86.4%62.5%
TriviaQA84.8%75.2%
ARC (AI2) Challenge—78.6%
BoolQ—87.4%
OpenBookQA—79.8%

Multilingual GPT-4 leads

GPT-4: 40.6 (#215), Mistral 7B: 25.8 (#283)

BenchmarkGPT-4Mistral 7B
LMArena Non-English12461012
LMArena Chinese12421009
LMArena French12831037
LMArena German1251987
LMArena Japanese1209878
LMArena Russian12511018
LMArena Spanish12611026
LMArena Korean1184—

Instruction Following GPT-4 leads

GPT-4: 65.3 (#222), Mistral 7B: 54.2 (#280)

BenchmarkGPT-4Mistral 7B
LMArena Instruction Following12411060

Long Context GPT-4 leads

GPT-4: 37.7 (#212), Mistral 7B: 32.2 (#271)

BenchmarkGPT-4Mistral 7B
LMArena Longer Query12441060

Writing & Preference GPT-4 leads

GPT-4: 34.9 (#268), Mistral 7B: 30.7 (#286)

BenchmarkGPT-4Mistral 7B
LMArena Text12631090
LMArena Creative Writing12441068
LMArena Multi-Turn12571062
EQ-Bench Creative Writing752—

Frequently asked questions

Is GPT-4 better than Mistral 7B?

GPT-4 is the stronger model overall, scoring 29.1 to 23.0 on the Noometry Index. Mistral 7B costs 150× less per token, which makes it the better buy when GPT-4's lead doesn't matter for your workload.

Which is cheaper, GPT-4 or Mistral 7B?

Mistral 7B is cheaper. It lists at $0.25 per million input tokens and $0.25 per million output tokens; GPT-4 lists at $30 and $60.

Is GPT-4 or Mistral 7B better for coding?

GPT-4 scores higher on coding benchmarks: 31.6 versus 26.4 in the Noometry coding category.

Which has the bigger context window?

GPT-4 does, with 8K tokens against 8K.

How many benchmarks do GPT-4 and Mistral 7B share?

31 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Mistral 7B has 37.

Related comparisons

Go deeper