Noometry

Model comparison

GPT-5.4 vs Llama 4 Maverick

GPT-5.4 is the stronger model overall, scoring 59.4 to 30.9 on the Noometry Index. Llama 4 Maverick costs 19× less per token, which makes it the better buy when GPT-5.4's lead doesn't matter for your workload.

Last verified . 37 shared benchmarks.

GPT-5.4 OpenAI

59.4

Rank #16 Confirmed

Llama 4 Maverick Meta

30.9

Rank #282 Confirmed

Summary

  • They share 37 benchmarks with published results for both. GPT-5.4 scores higher in 10 categories and Llama 4 Maverick in 0 categories; 10 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-5.4 leads 61.8 to 10.1.
  • The biggest single-benchmark swing is ARC-AGI-1: 93.7% for GPT-5.4 and 4.4% for Llama 4 Maverick.
  • Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $2.50 / $15 for GPT-5.4.
  • GPT-5.4 accepts more context: 1.05M tokens versus 128K.
  • Llama 4 Maverick has downloadable open weights; the other is API-only.

Side by side

GPT-5.4Llama 4 Maverick
ProviderOpenAIMeta
Noometry Index59.430.9
Released2026-03-052025-04-05
WeightsProprietaryOpen
Context window1.05M128K
Max output128K4K
Input $ / M tokens$2.50$0.19
Output $ / M tokens$15$0.65
Results tracked6854

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), Llama 4 Maverick: 26.6 (#324)

BenchmarkGPT-5.4Llama 4 Maverick
SciCode56.6%33.1%
WeirdML77.7%24.5%
LMArena Coding14971302
ALE-Bench1,607172.97
SWE-bench Verified76.9%—
DeepSWE51.8%—
SWE-bench Verified (bash only)—21%
Aider Polyglot—15.6%
LMArena WebDev1465—
GSO31.4%—
BigCodeBench Instruct—49.7%
MirrorCode15.6%—
BigCodeBench Complete—61.4%
AlgoTune1.85—

Agentic & Tool Use GPT-5.4 leads

GPT-5.4: 46.5 (#13), Llama 4 Maverick: 28.2 (#91)

BenchmarkGPT-5.4Llama 4 Maverick
Terminal-Bench81.8%—
APEX-Agents52.4%—
Berkeley Function Calling Leaderboard—37.3%
τ²-bench Banking39.4%—
DeepResearch Bench35.1%—
PostTrainBench19%—
GBAEval45.1%—
LMArena Search1197—
METR Time Horizons74.3%—
Vending-Bench 26,144—

Reasoning GPT-5.4 leads

GPT-5.4: 61.8 (#19), Llama 4 Maverick: 10.1 (#342)

BenchmarkGPT-5.4Llama 4 Maverick
ARC-AGI-274%0%
Kagi LLM Benchmark63.8%55.9%
NYT Connections (extended)91.3%8%
ARC-AGI-193.7%4.4%
CritPt23.4%0%
EnigmaEval16%0.6%
LMArena Hard Prompts14851281
DTBench94.4%61.9%
LMCA52%15.9%
Epoch Capabilities Index156.81132.2
ForecastBench59.557.5
SimpleBench—27.7%
Chess Puzzles44%—
Thematic Generalization80%—
EBR-Bench25.4%—
Mystery Game Puzzles37%—

Math GPT-5.4 leads

GPT-5.4: 73.5 (#19), Llama 4 Maverick: 26.0 (#262)

BenchmarkGPT-5.4Llama 4 Maverick
OTIS Mock AIME 2024-202597.8%20.6%
LMArena Math14881299
FrontierMath (Feb 2025 set)47.6%0.7%
FrontierMath (Tiers 1-3)78.6%—
FrontierMath Tier 449%—
MathArena Final-Answer Competitions83.1%—
ProofBench56%—
Omni-MATH—42.2%
MATH Level 5—73%
FrontierMath Tier 4 (v1)27.1%—

Knowledge GPT-5.4 leads

GPT-5.4: 65.3 (#14), Llama 4 Maverick: 33.4 (#204)

BenchmarkGPT-5.4Llama 4 Maverick
GPQA Diamond93.3%67%
Humanity's Last Exam36.2%5.7%
Vectara Hallucination Rate7%8.2%
LMArena Expert15071259
SimpleQA Verified45.1%—
MMLU-Pro—81%
Confabulations—22.6%
GPQA (HELM)—65%

Multimodal GPT-5.4 leads

GPT-5.4: 43.7 (#20), Llama 4 Maverick: 31.6 (#105)

BenchmarkGPT-5.4Llama 4 Maverick
LMArena Vision13031142
GeoBench—52%
Blueprint-Bench 227.1%—
Furniture Assembly37.5%—
LMArena Document1471—
SpatialViz-Bench—31.8%

Multilingual GPT-5.4 leads

GPT-5.4: 56.2 (#23), Llama 4 Maverick: 42.2 (#195)

BenchmarkGPT-5.4Llama 4 Maverick
LMArena Non-English14651269
LMArena Chinese15191277
LMArena French14931259
LMArena German14721291
LMArena Japanese14851207
LMArena Korean14481203
LMArena Russian14801286
LMArena Spanish14541293

Instruction Following GPT-5.4 leads

GPT-5.4: 77.1 (#27), Llama 4 Maverick: 71.7 (#146)

BenchmarkGPT-5.4Llama 4 Maverick
LMArena Instruction Following14691267
IFEval—90.8%

Long Context GPT-5.4 leads

GPT-5.4: 50.3 (#8), Llama 4 Maverick: 31.4 (#279)

BenchmarkGPT-5.4Llama 4 Maverick
LMArena Longer Query14731280
Fiction.LiveBench—46.2%
CL-bench27.9%—
CL-bench Life21.7%—

Writing & Preference GPT-5.4 leads

GPT-5.4: 71.9 (#17), Llama 4 Maverick: 38.8 (#252)

BenchmarkGPT-5.4Llama 4 Maverick
LMArena Text14691287
LMArena Creative Writing14391267
EQ-Bench Creative Writing1840860
LMArena Multi-Turn14821289
Short-Story Creative Writing—62%
WildBench—80%
EQ-Bench 41272—

Frequently asked questions

Is GPT-5.4 better than Llama 4 Maverick?

GPT-5.4 is the stronger model overall, scoring 59.4 to 30.9 on the Noometry Index. Llama 4 Maverick costs 19× 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 Llama 4 Maverick?

Llama 4 Maverick is cheaper. It lists at $0.19 per million input tokens and $0.65 per million output tokens; GPT-5.4 lists at $2.50 and $15.

Is GPT-5.4 or Llama 4 Maverick better for coding?

GPT-5.4 scores higher on coding benchmarks: 52.6 versus 26.6 in the Noometry coding category.

Which has the bigger context window?

GPT-5.4 does, with 1.05M tokens against 128K.

How many benchmarks do GPT-5.4 and Llama 4 Maverick share?

37 benchmarks have published results for both models. GPT-5.4 has 68 scored results on Noometry and Llama 4 Maverick has 54.

Related comparisons

Go deeper