Noometry

Model comparison

GPT-6 Astra vs Llama 4 Maverick

GPT-6 Astra is the stronger model overall, scoring 70.3 to 30.6 on the Noometry Index. Llama 4 Maverick costs 66× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.

Last verified . 30 shared benchmarks.

GPT-6 Astra OpenAI

70.3

Rank #1 Confirmed

Llama 4 Maverick Meta

30.6

Rank #276 Confirmed

Summary

  • They share 30 benchmarks with published results for both. GPT-6 Astra 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-6 Astra leads 85.1 to 10.1.
  • The biggest single-benchmark swing is ARC-AGI-2: 95% for GPT-6 Astra and 0% for Llama 4 Maverick.
  • Llama 4 Maverick is cheaper at $0.19 / $0.65 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
  • GPT-6 Astra accepts more context: 1.05M tokens versus 128K.
  • Llama 4 Maverick has downloadable open weights; the other is API-only.

Side by side

GPT-6 AstraLlama 4 Maverick
ProviderOpenAIMeta
Noometry Index70.330.6
Released2026-09-032025-04-05
WeightsProprietaryOpen
Context window1.05M128K
Max output128K4K
Input $ / M tokens$10$0.19
Output $ / M tokens$50$0.65
Results tracked5154

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

Category by category

Coding GPT-6 Astra leads

GPT-6 Astra: 70.5 (#3), Llama 4 Maverick: 26.3 (#318)

BenchmarkGPT-6 AstraLlama 4 Maverick
SciCode56.5%33.1%
WeirdML93.6%24.5%
LMArena Coding14881302
ALE-Bench2,951172.97
DeepSWE74.1%—
FrontierCode53.3%—
SWE-bench Verified (bash only)—21%
Aider Polyglot—15.6%
LMArena WebDev1792—
FrontierSWE65.5%—
BigCodeBench Instruct—49.7%
MirrorCode46.7%—
BigCodeBench Complete—61.4%

Agentic & Tool Use GPT-6 Astra leads

GPT-6 Astra: 52.3 (#3), Llama 4 Maverick: 27.9 (#86)

BenchmarkGPT-6 AstraLlama 4 Maverick
APEX-Agents64.7%—
Berkeley Function Calling Leaderboard—37.3%
Remote Labor Index20.8%—
BALROG68.3%—
Vending-Bench 215,515—

Reasoning GPT-6 Astra leads

GPT-6 Astra: 85.1 (#1), Llama 4 Maverick: 10.1 (#337)

BenchmarkGPT-6 AstraLlama 4 Maverick
ARC-AGI-295%0%
NYT Connections (extended)98.1%8%
ARC-AGI-198.5%4.4%
CritPt31.7%0%
LMArena Hard Prompts14641281
DTBench97.3%61.9%
LMCA64.4%15.9%
Epoch Capabilities Index166.6132.2
SimpleBench—27.7%
Kagi LLM Benchmark—55.9%
Chess Puzzles72%—
EnigmaEval—0.6%
EBR-Bench76.2%—
Mystery Game Puzzles84%—
Bench to the Future 30.14—
ForecastBench—57.5

Math GPT-6 Astra leads

GPT-6 Astra: 93.4 (#1), Llama 4 Maverick: 24.6 (#256)

BenchmarkGPT-6 AstraLlama 4 Maverick
OTIS Mock AIME 2024-2025100%20.6%
LMArena Math14631299
FrontierMath (Tiers 1-3)93.7%—
FrontierMath Tier 497.6%—
ProofBench99%—
Omni-MATH—42.2%
MATH Level 5—73%
FrontierMath (Feb 2025 set)—0.7%
FrontierMath Erdős2.9%—

Knowledge GPT-6 Astra leads

GPT-6 Astra: 75.3 (#1), Llama 4 Maverick: 33.3 (#196)

BenchmarkGPT-6 AstraLlama 4 Maverick
GPQA Diamond95.8%67%
Humanity's Last Exam54.8%5.7%
Vectara Hallucination Rate8.7%8.2%
LMArena Expert14861259
SimpleQA Verified75.6%—
MMLU-Pro—81%
Confabulations—22.6%
GPQA (HELM)—65%

Multimodal GPT-6 Astra leads

GPT-6 Astra: 55.1 (#3), Llama 4 Maverick: 30.9 (#100)

BenchmarkGPT-6 AstraLlama 4 Maverick
LMArena Vision12801141
GeoBench—52%
Blueprint-Bench 249.7%—
Furniture Assembly80%—
LMArena Document1468—
SpatialViz-Bench—31.8%

Multilingual GPT-6 Astra leads

GPT-6 Astra: 54.0 (#53), Llama 4 Maverick: 42.4 (#190)

BenchmarkGPT-6 AstraLlama 4 Maverick
LMArena Non-English14311269
LMArena Chinese14951277
LMArena French14671258
LMArena Russian14341286
LMArena Spanish14121293
LMArena German—1291
LMArena Japanese—1207
LMArena Korean—1203

Instruction Following GPT-6 Astra leads

GPT-6 Astra: 76.6 (#37), Llama 4 Maverick: 71.9 (#140)

BenchmarkGPT-6 AstraLlama 4 Maverick
LMArena Instruction Following14541267
IFEval—90.8%

Long Context GPT-6 Astra leads

GPT-6 Astra: 44.7 (#57), Llama 4 Maverick: 31.4 (#274)

BenchmarkGPT-6 AstraLlama 4 Maverick
LMArena Longer Query14601280
Fiction.LiveBench—46.2%

Writing & Preference GPT-6 Astra leads

GPT-6 Astra: 75.7 (#7), Llama 4 Maverick: 39.0 (#247)

BenchmarkGPT-6 AstraLlama 4 Maverick
LMArena Text14431287
LMArena Creative Writing14231266
EQ-Bench Creative Writing2173860
LMArena Multi-Turn14571289
Short-Story Creative Writing—62%
WildBench—80%

Frequently asked questions

Is GPT-6 Astra better than Llama 4 Maverick?

GPT-6 Astra is the stronger model overall, scoring 70.3 to 30.6 on the Noometry Index. Llama 4 Maverick costs 66× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.

Which is cheaper, GPT-6 Astra 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-6 Astra lists at $10 and $50.

Is GPT-6 Astra or Llama 4 Maverick better for coding?

GPT-6 Astra scores higher on coding benchmarks: 70.5 versus 26.3 in the Noometry coding category.

Which has the bigger context window?

GPT-6 Astra does, with 1.05M tokens against 128K.

How many benchmarks do GPT-6 Astra and Llama 4 Maverick share?

30 benchmarks have published results for both models. GPT-6 Astra has 51 scored results on Noometry and Llama 4 Maverick has 54.

Related comparisons

Go deeper