Noometry

Model comparison

GPT-5.4 vs Kimi K2 (Jul 2025)

GPT-5.4 is the stronger model overall, scoring 59.4 to 41.2 on the Noometry Index. Kimi K2 (Jul 2025) costs 5.6× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

Last verified . 30 shared benchmarks.

GPT-5.4 OpenAI

59.4

Rank #16 Confirmed

Kimi K2 (Jul 2025) Moonshot AI

41.2

Rank #140 Confirmed

Summary

  • They share 30 benchmarks with published results for both. GPT-5.4 scores higher in 9 categories and Kimi K2 (Jul 2025) in 0 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-5.4 leads 61.8 to 23.3.
  • The biggest single-benchmark swing is Terminal-Bench: 81.8% for GPT-5.4 and 35.7% for Kimi K2 (Jul 2025).
  • Kimi K2 (Jul 2025) is cheaper at $0.57 / $2.30 per million input/output tokens, against $2.50 / $15 for GPT-5.4.
  • GPT-5.4 accepts more context: 1.05M tokens versus 262K.
  • Kimi K2 (Jul 2025) has downloadable open weights; the other is API-only.

Side by side

GPT-5.4Kimi K2 (Jul 2025)
ProviderOpenAIMoonshot AI
Noometry Index59.441.2
Released2026-03-052025-07-12
WeightsProprietaryOpen
Context window1.05M262K
Max output128K262K
Input $ / M tokens$2.50$0.57
Output $ / M tokens$15$2.30
Results tracked6842

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

Category by category

Coding GPT-5.4 leads

GPT-5.4: 52.6 (#33), Kimi K2 (Jul 2025): 42.4 (#102)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
GSO31.4%4.9%
WeirdML77.7%42.8%
LMArena Coding14971399
ALE-Bench1,607597.5
SWE-bench Verified76.9%—
DeepSWE51.8%—
SWE-bench Verified (bash only)—63.4%
Aider Polyglot—59.1%
LMArena WebDev1465—
SciCode56.6%—
MirrorCode15.6%—
AlgoTune1.85—

Agentic & Tool Use GPT-5.4 leads

GPT-5.4: 46.5 (#13), Kimi K2 (Jul 2025): 32.4 (#64)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
Terminal-Bench81.8%35.7%
METR Time Horizons74.3%59.2%
APEX-Agents52.4%—
Berkeley Function Calling Leaderboard—59.1%
τ²-bench Banking39.4%—
DeepResearch Bench35.1%—
PostTrainBench19%—
GBAEval45.1%—
LMArena Search1197—
Vending-Bench 26,144—

Reasoning GPT-5.4 leads

GPT-5.4: 61.8 (#19), Kimi K2 (Jul 2025): 23.3 (#179)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
Kagi LLM Benchmark63.8%64.4%
LMArena Hard Prompts14851384
Epoch Capabilities Index156.81146.01
ForecastBench59.560.2
ARC-AGI-274%—
SimpleBench—26.3%
NYT Connections (extended)91.3%—
ARC-AGI-193.7%—
CritPt23.4%—
Chess Puzzles44%—
EnigmaEval16%—
Thematic Generalization80%—
EBR-Bench25.4%—
Mystery Game Puzzles37%—
DTBench94.4%—
LMCA52%—

Math GPT-5.4 leads

GPT-5.4: 73.5 (#19), Kimi K2 (Jul 2025): 42.7 (#83)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
LMArena Math14881397
FrontierMath (Feb 2025 set)47.6%21.4%
FrontierMath Tier 4 (v1)27.1%0%
FrontierMath (Tiers 1-3)78.6%—
FrontierMath Tier 449%—
MathArena Final-Answer Competitions83.1%—
OTIS Mock AIME 2024-202597.8%—
ProofBench56%—
Omni-MATH—65.4%

Knowledge GPT-5.4 leads

GPT-5.4: 65.3 (#14), Kimi K2 (Jul 2025): 37.3 (#157)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
Vectara Hallucination Rate7%17.9%
LMArena Expert15071365
GPQA Diamond93.3%—
Humanity's Last Exam36.2%—
SimpleQA Verified45.1%—
MMLU-Pro—81.9%
Confabulations—20.4%
GPQA (HELM)—65.3%

Multimodal Not comparable

GPT-5.4: 43.7 (#20), Kimi K2 (Jul 2025): —

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
LMArena Vision1303—
Blueprint-Bench 227.1%—
Furniture Assembly37.5%—
LMArena Document1471—

Multilingual GPT-5.4 leads

GPT-5.4: 56.2 (#23), Kimi K2 (Jul 2025): 49.6 (#130)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
LMArena Non-English14651372
LMArena Chinese15191415
LMArena French14931379
LMArena German14721387
LMArena Japanese14851349
LMArena Korean14481325
LMArena Russian14801385
LMArena Spanish14541386

Instruction Following GPT-5.4 leads

GPT-5.4: 77.1 (#27), Kimi K2 (Jul 2025): 71.1 (#156)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
LMArena Instruction Following14691348
IFEval—85%

Long Context GPT-5.4 leads

GPT-5.4: 50.3 (#8), Kimi K2 (Jul 2025): 41.2 (#145)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
CL-bench27.9%17.6%
LMArena Longer Query14731353
Fiction.LiveBench—66.7%
CL-bench Life21.7%—

Writing & Preference GPT-5.4 leads

GPT-5.4: 71.9 (#17), Kimi K2 (Jul 2025): 62.3 (#78)

BenchmarkGPT-5.4Kimi K2 (Jul 2025)
LMArena Text14691380
LMArena Creative Writing14391350
EQ-Bench Creative Writing18401666
LMArena Multi-Turn14821371
Short-Story Creative Writing—85.6%
WildBench—86.2%
EQ-Bench 41272—

Frequently asked questions

Is GPT-5.4 better than Kimi K2 (Jul 2025)?

GPT-5.4 is the stronger model overall, scoring 59.4 to 41.2 on the Noometry Index. Kimi K2 (Jul 2025) costs 5.6× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

Which is cheaper, GPT-5.4 or Kimi K2 (Jul 2025)?

Kimi K2 (Jul 2025) is cheaper. It lists at $0.57 per million input tokens and $2.30 per million output tokens; GPT-5.4 lists at $2.50 and $15.

Is GPT-5.4 or Kimi K2 (Jul 2025) better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.6 versus 42.4 in the Noometry coding category.

Which has the bigger context window?

GPT-5.4 does, with 1.05M tokens against 262K.

How many benchmarks do GPT-5.4 and Kimi K2 (Jul 2025) share?

30 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and Kimi K2 (Jul 2025) has 42.

Related comparisons

Go deeper