Noometry

Model comparison

Claude Sonnet 4.5 vs GPT-6 Astra

GPT-6 Astra is the stronger model overall, scoring 70.8 to 44.1 on the Noometry Index. Claude Sonnet 4.5 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 . 44 shared benchmarks.

Claude Sonnet 4.5 Anthropic

44.1

Rank #81 Confirmed

GPT-6 Astra OpenAI

70.8

Rank #1 Confirmed

Summary

  • They share 44 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 1 category and GPT-6 Astra in 9 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where GPT-6 Astra leads 93.5 to 32.3.
  • The biggest single-benchmark swing is FrontierMath Tier 4: 2.4% for Claude Sonnet 4.5 and 97.6% for GPT-6 Astra.
  • Claude Sonnet 4.5 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 4.5GPT-6 Astra
ProviderAnthropicOpenAI
Noometry Index44.170.8
Released2025-09-292026-09-03
WeightsProprietaryProprietary
Context window200K1.05M
Max output64K128K
Input $ / M tokens$3$10
Output $ / M tokens$15$50
Results tracked7356

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

Category by category

Coding GPT-6 Astra leads

Claude Sonnet 4.5: 47.3 (#61), GPT-6 Astra: 73.7 (#2)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena WebDev13931786
SciCode44.7%56.5%
GSO14.7%79.4%
WeirdML47.7%93.6%
LMArena Coding14891487
ALE-Bench796.152,951
SWE-bench Verified71.3%—
DeepSWE—74.1%
FrontierCode—53.3%
SWE-bench Verified (bash only)71.4%—
SWE-bench Multilingual67%—
FrontierSWE—65.5%
MirrorCode—46.7%
AlgoTune1.52—

Agentic & Tool Use GPT-6 Astra leads

Claude Sonnet 4.5: 38.3 (#32), GPT-6 Astra: 52.9 (#3)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
Remote Labor Index2.1%20.8%
Vending-Bench 23,83915,515
Terminal-Bench46.5%—
APEX-Agents—64.7%
Berkeley Function Calling Leaderboard73.2%—
GDPval42.5%—
τ²-bench Airline72%—
τ²-bench Banking25.3%—
τ²-bench Retail72.4%—
τ²-bench Telecom84.9%—
Cybench60%—
DeepResearch Bench52.6%—
OSWorld62.9%—
BALROG—68.3%
GDP.pdf—34.2%
LMArena Search1159—
METR Time Horizons67.4%—

Reasoning GPT-6 Astra leads

Claude Sonnet 4.5: 26.9 (#125), GPT-6 Astra: 85.1 (#1)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
ARC-AGI-213.6%95%
NYT Connections (extended)37.3%98.1%
ARC-AGI-163.7%98.5%
CritPt1.1%31.7%
Chess Puzzles12%72%
EBR-Bench2.4%76.2%
LMArena Hard Prompts14621462
Mystery Game Puzzles17%84%
DTBench83.2%97.3%
LMCA38.8%64.4%
Epoch Capabilities Index146.84166.45
SimpleBench54.3%—
Kagi LLM Benchmark57.9%—
EnigmaEval6%—
Bench to the Future 3—0.14
ForecastBench61.9—

Math GPT-6 Astra leads

Claude Sonnet 4.5: 32.3 (#216), GPT-6 Astra: 93.5 (#2)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
FrontierMath (Tiers 1-3)23.9%93.7%
FrontierMath Tier 42.4%97.6%
OTIS Mock AIME 2024-202577.8%100%
ProofBench19%99%
LMArena Math14491465
Omni-MATH55.3%—
MATH Level 597.7%—
FrontierMath (Feb 2025 set)15.2%—
FrontierMath Erdős—2.9%
FrontierMath Tier 4 (v1)4.2%—

Knowledge GPT-6 Astra leads

Claude Sonnet 4.5: 48.4 (#76), GPT-6 Astra: 75.3 (#1)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
GPQA Diamond82.3%95.8%
Humanity's Last Exam13.7%54.8%
SimpleQA Verified30.7%75.6%
Vectara Hallucination Rate12%8.7%
LMArena Expert14821483
MMLU-Pro86.9%—
GPQA (HELM)68.6%—

Multimodal GPT-6 Astra leads

Claude Sonnet 4.5: 34.8 (#89), GPT-6 Astra: 55.0 (#3)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena Document14501468
LMArena Vision—1281
VPCT39.8%—
Blueprint-Bench 2—49.7%
Furniture Assembly—80%

Multilingual Too close to call

Claude Sonnet 4.5: 53.4 (#69), GPT-6 Astra: 53.7 (#61)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena Non-English14251430
LMArena Chinese14591484
LMArena French14581456
LMArena German14271440
LMArena Japanese13901379
LMArena Korean14031426
LMArena Russian14371436
LMArena Spanish14571407

Instruction Following GPT-6 Astra leads

Claude Sonnet 4.5: 75.0 (#78), GPT-6 Astra: 76.3 (#44)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena Instruction Following14591450
IFEval85%—

Long Context Too close to call

Claude Sonnet 4.5: 45.2 (#46), GPT-6 Astra: 44.5 (#62)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena Longer Query14761456

Writing & Preference GPT-6 Astra leads

Claude Sonnet 4.5: 66.5 (#34), GPT-6 Astra: 75.3 (#7)

BenchmarkClaude Sonnet 4.5GPT-6 Astra
LMArena Text14391441
LMArena Creative Writing14421418
EQ-Bench Creative Writing16782173
LMArena Multi-Turn14651448
WildBench85.4%—

Frequently asked questions

Is Claude Sonnet 4.5 better than GPT-6 Astra?

GPT-6 Astra is the stronger model overall, scoring 70.8 to 44.1 on the Noometry Index. Claude Sonnet 4.5 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 4.5 or GPT-6 Astra?

Claude Sonnet 4.5 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 4.5 or GPT-6 Astra better for coding?

GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 47.3 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 4.5 and GPT-6 Astra share?

44 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and GPT-6 Astra has 56.

Related comparisons

Go deeper