Model comparison

GPT-4 vs Kimi K2.5 Instant

Kimi K2.5 Instant is the stronger model overall, scoring 43.6 to 29.1 on the Noometry Index.

Last verified . 16 shared benchmarks.

GPT-4 OpenAI

29.1

Rank #316 Confirmed

Kimi K2.5 Instant Moonshot AI

43.6

Rank #89 Confirmed

Summary

  • They share 16 benchmarks with published results for both. GPT-4 scores higher in 0 categories and Kimi K2.5 Instant in 8 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Kimi K2.5 Instant leads 39.4 to 10.8.
  • Kimi K2.5 Instant has downloadable open weights; the other is API-only.

Side by side

GPT-4 and Kimi K2.5 Instant specifications
GPT-4Kimi K2.5 Instant
ProviderOpenAIMoonshot AI
Noometry Index29.143.6
Released2023-03-14—
WeightsProprietaryOpen
Context window8K—
Max output8K—
Input $ / M tokens$30—
Output $ / M tokens$60—
Results tracked3818

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

Category by category

Coding Kimi K2.5 Instant leads

GPT-4: 31.6 (#283), Kimi K2.5 Instant: 42.6 (#97)

Coding benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Coding12541484
LMArena WebDev—1404
WeirdML12.4%—
BigCodeBench Instruct46%—
BigCodeBench Complete57.2%—
HumanEval+79.3%—

Agentic & Tool Use Not comparable

GPT-4: —, Kimi K2.5 Instant: —

Agentic & Tool Use benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
METR Time Horizons36.1%—

Reasoning Kimi K2.5 Instant leads

GPT-4: 17.8 (#289), Kimi K2.5 Instant: 29.7 (#90)

Reasoning benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Hard Prompts12411443
Chess Puzzles4%—
Mystery Game Puzzles12%—
DTBench62.7%—
LMCA17.1%—
BIG-Bench Hard75.1%—
Epoch Capabilities Index125.89—
ForecastBench57.8—
HellaSwag95.3%—
WinoGrande87.5%—

Math Kimi K2.5 Instant leads

GPT-4: 10.8 (#309), Kimi K2.5 Instant: 39.4 (#105)

Math benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Math12691442
OTIS Mock AIME 2024-20251.1%—
MATH Level 523%—
GSM8K92%—

Knowledge Kimi K2.5 Instant leads

GPT-4: 18.4 (#282), Kimi K2.5 Instant: 40.2 (#123)

Knowledge benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Expert12111440
GPQA Diamond35.7%—
MMLU86.4%—
TriviaQA84.8%—

Multimodal Not comparable

GPT-4: —, Kimi K2.5 Instant: 40.2 (#50)

Multimodal benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Vision—1254

Multilingual Kimi K2.5 Instant leads

GPT-4: 40.6 (#215), Kimi K2.5 Instant: 52.0 (#94)

Multilingual benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Non-English12461406
LMArena Chinese12421449
LMArena French12831403
LMArena German12511413
LMArena Korean11841378
LMArena Russian12511404
LMArena Spanish12611447
LMArena Japanese1209—

Instruction Following Kimi K2.5 Instant leads

GPT-4: 65.3 (#222), Kimi K2.5 Instant: 75.3 (#65)

Instruction Following benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Instruction Following12411430

Long Context Kimi K2.5 Instant leads

GPT-4: 37.7 (#212), Kimi K2.5 Instant: 43.9 (#83)

Long Context benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Longer Query12441435

Writing & Preference Kimi K2.5 Instant leads

GPT-4: 34.9 (#268), Kimi K2.5 Instant: 60.6 (#95)

Writing & Preference benchmarks
BenchmarkGPT-4Kimi K2.5 Instant
LMArena Text12631420
LMArena Creative Writing12441381
LMArena Multi-Turn12571427
EQ-Bench Creative Writing752—

Frequently asked questions

Is GPT-4 better than Kimi K2.5 Instant?

Kimi K2.5 Instant is the stronger model overall, scoring 43.6 to 29.1 on the Noometry Index.

Is GPT-4 or Kimi K2.5 Instant better for coding?

Kimi K2.5 Instant scores higher on coding benchmarks: 42.6 versus 31.6 in the Noometry coding category.

How many benchmarks do GPT-4 and Kimi K2.5 Instant share?

16 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Kimi K2.5 Instant has 18.

Related comparisons

Go deeper