Noometry

Model comparison

Claude Sonnet 4 vs GPT-6 Astra

GPT-6 Astra is the stronger model overall, scoring 70.3 to 40.6 on the Noometry Index. Claude Sonnet 4 costs 3.3× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.

Last verified . 29 shared benchmarks.

Claude Sonnet 4 Anthropic

40.6

Rank #139 Confirmed

GPT-6 Astra OpenAI

70.3

Rank #1 Confirmed

Summary

  • They share 29 benchmarks with published results for both. Claude Sonnet 4 scores higher in 0 categories and GPT-6 Astra in 10 categories; 10 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 23.0.
  • The biggest single-benchmark swing is ARC-AGI-2: 5.9% for Claude Sonnet 4 and 95% for GPT-6 Astra.
  • Claude Sonnet 4 is cheaper at $3 / $15 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
  • GPT-6 Astra accepts more context: 1.05M tokens versus 200K.

Side by side

Claude Sonnet 4GPT-6 Astra
ProviderAnthropicOpenAI
Noometry Index40.670.3
Released2025-05-222026-09-03
WeightsProprietaryProprietary
Context window200K1.05M
Max output64K128K
Input $ / M tokens$3$10
Output $ / M tokens$15$50
Results tracked5851

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

Category by category

Coding GPT-6 Astra leads

Claude Sonnet 4: 44.0 (#70), GPT-6 Astra: 70.5 (#3)

BenchmarkClaude Sonnet 4GPT-6 Astra
SciCode40%56.5%
WeirdML46.1%93.6%
LMArena Coding14141488
ALE-Bench655.352,951
DeepSWE—74.1%
FrontierCode—53.3%
SWE-bench Verified (bash only)64.9%—
Aider Polyglot61.3%—
LMArena WebDev—1792
FrontierSWE—65.5%
GSO4.9%—
MirrorCode—46.7%

Agentic & Tool Use GPT-6 Astra leads

Claude Sonnet 4: 38.3 (#29), GPT-6 Astra: 52.3 (#3)

BenchmarkClaude Sonnet 4GPT-6 Astra
APEX-Agents—64.7%
Remote Labor Index—20.8%
TheAgentCompany33.1%—
Cybench35%—
DeepResearch Bench46.6%—
OSWorld43.9%—
BALROG—68.3%
METR Time Horizons62%—
Vending-Bench 2—15,515

Reasoning GPT-6 Astra leads

Claude Sonnet 4: 23.0 (#169), GPT-6 Astra: 85.1 (#1)

BenchmarkClaude Sonnet 4GPT-6 Astra
ARC-AGI-25.9%95%
ARC-AGI-140%98.5%
CritPt0.3%31.7%
LMArena Hard Prompts13721464
DTBench77.1%97.3%
LMCA29%64.4%
Epoch Capabilities Index141.69166.6
SimpleBench45.5%—
Kagi LLM Benchmark73%—
NYT Connections (extended)—98.1%
Chess Puzzles—72%
EnigmaEval3.1%—
EBR-Bench—76.2%
Mystery Game Puzzles—84%
Bench to the Future 3—0.14
ForecastBench60.2—

Math GPT-6 Astra leads

Claude Sonnet 4: 41.5 (#71), GPT-6 Astra: 93.4 (#1)

BenchmarkClaude Sonnet 4GPT-6 Astra
OTIS Mock AIME 2024-202571.1%100%
LMArena Math13751463
FrontierMath (Tiers 1-3)—93.7%
FrontierMath Tier 4—97.6%
ProofBench—99%
Omni-MATH60.2%—
MATH Level 584.4%—
FrontierMath (Feb 2025 set)4.1%—
FrontierMath Erdős—2.9%
FrontierMath Tier 4 (v1)0%—

Knowledge GPT-6 Astra leads

Claude Sonnet 4: 41.6 (#99), GPT-6 Astra: 75.3 (#1)

BenchmarkClaude Sonnet 4GPT-6 Astra
GPQA Diamond79.2%95.8%
Humanity's Last Exam7.8%54.8%
Vectara Hallucination Rate10.3%8.7%
LMArena Expert13711486
SimpleQA Verified—75.6%
MMLU-Pro84.3%—
Confabulations13.2%—
GPQA (HELM)70.6%—

Multimodal GPT-6 Astra leads

Claude Sonnet 4: 25.6 (#116), GPT-6 Astra: 55.1 (#3)

BenchmarkClaude Sonnet 4GPT-6 Astra
LMArena Vision11911280
GeoBench37%—
VPCT34%—
Blueprint-Bench 2—49.7%
Furniture Assembly—80%
LMArena Document—1468
MindCube44.8%—

Multilingual GPT-6 Astra leads

Claude Sonnet 4: 46.9 (#151), GPT-6 Astra: 54.0 (#53)

BenchmarkClaude Sonnet 4GPT-6 Astra
LMArena Non-English13331431
LMArena Chinese13511495
LMArena French13641467
LMArena Russian13551434
LMArena Spanish13561412
LMArena German1331—
LMArena Japanese1301—
LMArena Korean1293—

Instruction Following GPT-6 Astra leads

Claude Sonnet 4: 71.8 (#141), GPT-6 Astra: 76.6 (#37)

BenchmarkClaude Sonnet 4GPT-6 Astra
LMArena Instruction Following13751454
IFEval84%—

Long Context GPT-6 Astra leads

Claude Sonnet 4: 33.6 (#254), GPT-6 Astra: 44.7 (#57)

BenchmarkClaude Sonnet 4GPT-6 Astra
LMArena Longer Query13981460
Fiction.LiveBench46.9%—

Writing & Preference GPT-6 Astra leads

Claude Sonnet 4: 57.3 (#127), GPT-6 Astra: 75.7 (#7)

BenchmarkClaude Sonnet 4GPT-6 Astra
LMArena Text13511443
LMArena Creative Writing13451423
EQ-Bench Creative Writing14832173
LMArena Multi-Turn13761457
Short-Story Creative Writing81.4%—
WildBench83.8%—

Frequently asked questions

Is Claude Sonnet 4 better than GPT-6 Astra?

GPT-6 Astra is the stronger model overall, scoring 70.3 to 40.6 on the Noometry Index. Claude Sonnet 4 costs 3.3× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.

Which is cheaper, Claude Sonnet 4 or GPT-6 Astra?

Claude Sonnet 4 is cheaper. It lists at $3 per million input tokens and $15 per million output tokens; GPT-6 Astra lists at $10 and $50.

Is Claude Sonnet 4 or GPT-6 Astra better for coding?

GPT-6 Astra scores higher on coding benchmarks: 70.5 versus 44.0 in the Noometry coding category.

Which has the bigger context window?

GPT-6 Astra does, with 1.05M tokens against 200K.

How many benchmarks do Claude Sonnet 4 and GPT-6 Astra share?

29 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and GPT-6 Astra has 51.

Related comparisons

Go deeper