Noometry

Model comparison

Claude Sonnet 4.5 vs DeepSeek V4 Flash

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 44.1 on the Noometry Index.

Last verified . 39 shared benchmarks.

Claude Sonnet 4.5 Anthropic

44.1

Rank #81 Confirmed

DeepSeek V4 Flash DeepSeek

53.6

Rank #35 Confirmed

Summary

  • They share 39 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 4 categories and DeepSeek V4 Flash in 4 categories; 5 gaps are clear of the uncertainty.
  • The widest gap is in math, where DeepSeek V4 Flash leads 60.3 to 32.3.
  • The biggest single-benchmark swing is NYT Connections (extended): 37.3% for Claude Sonnet 4.5 and 89.6% for DeepSeek V4 Flash.
  • DeepSeek V4 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.5.
  • DeepSeek V4 Flash accepts more context: 1M tokens versus 200K.
  • DeepSeek V4 Flash has downloadable open weights; the other is API-only.

Side by side

Claude Sonnet 4.5DeepSeek V4 Flash
ProviderAnthropicDeepSeek
Noometry Index44.153.6
Released2025-09-292026-04-24
WeightsProprietaryOpen
Context window200K1M
Max output64K393K
Input $ / M tokens$3$0.15
Output $ / M tokens$15$0.60
Results tracked7341

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

Category by category

Coding Too close to call

Claude Sonnet 4.5: 47.3 (#61), DeepSeek V4 Flash: 47.9 (#59)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
LMArena WebDev13931582
SciCode44.7%49.9%
WeirdML47.7%63%
LMArena Coding14891457
ALE-Bench796.151,306
SWE-bench Verified71.3%—
FrontierCode—18.8%
SWE-bench Verified (bash only)71.4%—
SWE-bench Multilingual67%—
GSO14.7%—
AlgoTune1.52—

Agentic & Tool Use Not comparable

Claude Sonnet 4.5: 38.3 (#32), DeepSeek V4 Flash: —

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
Terminal-Bench46.5%—
Berkeley Function Calling Leaderboard73.2%—
GDPval42.5%—
Remote Labor Index2.1%—
τ²-bench Airline72%—
τ²-bench Banking25.3%—
τ²-bench Retail72.4%—
τ²-bench Telecom84.9%—
Cybench60%—
DeepResearch Bench52.6%—
OSWorld62.9%—
LMArena Search1159—
METR Time Horizons67.4%—
Vending-Bench 23,839—

Reasoning DeepSeek V4 Flash leads

Claude Sonnet 4.5: 26.9 (#125), DeepSeek V4 Flash: 53.7 (#30)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
ARC-AGI-213.6%61.4%
SimpleBench54.3%61.1%
Kagi LLM Benchmark57.9%52.2%
NYT Connections (extended)37.3%89.6%
ARC-AGI-163.7%89%
CritPt1.1%16.6%
Chess Puzzles12%33%
LMArena Hard Prompts14621444
Mystery Game Puzzles17%34%
DTBench83.2%90.9%
LMCA38.8%41.7%
Epoch Capabilities Index146.84154.49
EnigmaEval6%—
EBR-Bench2.4%—
ForecastBench61.9—

Math DeepSeek V4 Flash leads

Claude Sonnet 4.5: 32.3 (#216), DeepSeek V4 Flash: 60.3 (#37)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
FrontierMath (Tiers 1-3)23.9%57.5%
FrontierMath Tier 42.4%24.4%
OTIS Mock AIME 2024-202577.8%94.4%
ProofBench19%56%
LMArena Math14491427
MathArena Final-Answer Competitions—76.5%
Omni-MATH55.3%—
MATH Level 597.7%—
FrontierMath (Feb 2025 set)15.2%—
FrontierMath Tier 4 (v1)4.2%—

Knowledge DeepSeek V4 Flash leads

Claude Sonnet 4.5: 48.4 (#76), DeepSeek V4 Flash: 55.4 (#48)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
GPQA Diamond82.3%91%
SimpleQA Verified30.7%33.6%
LMArena Expert14821441
Humanity's Last Exam13.7%—
MMLU-Pro86.9%—
Vectara Hallucination Rate12%—
GPQA (HELM)68.6%—

Multimodal Not comparable

Claude Sonnet 4.5: 34.8 (#89), DeepSeek V4 Flash: —

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
VPCT39.8%—
LMArena Document1450—

Multilingual Too close to call

Claude Sonnet 4.5: 53.4 (#69), DeepSeek V4 Flash: 53.0 (#72)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
LMArena Non-English14251420
LMArena Chinese14591468
LMArena French14581439
LMArena German14271418
LMArena Japanese13901406
LMArena Korean14031384
LMArena Russian14371428
LMArena Spanish14571436

Instruction Following Too close to call

Claude Sonnet 4.5: 75.0 (#78), DeepSeek V4 Flash: 74.9 (#81)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
LMArena Instruction Following14591421
IFEval85%—

Long Context Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 45.2 (#46), DeepSeek V4 Flash: 43.8 (#85)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
LMArena Longer Query14761434

Writing & Preference Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 66.5 (#34), DeepSeek V4 Flash: 63.8 (#61)

BenchmarkClaude Sonnet 4.5DeepSeek V4 Flash
LMArena Text14391432
LMArena Creative Writing14421403
EQ-Bench Creative Writing16781559
LMArena Multi-Turn14651449
WildBench85.4%—

Frequently asked questions

Is Claude Sonnet 4.5 better than DeepSeek V4 Flash?

DeepSeek V4 Flash is the stronger model overall, scoring 53.6 to 44.1 on the Noometry Index.

Which is cheaper, Claude Sonnet 4.5 or DeepSeek V4 Flash?

DeepSeek V4 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; Claude Sonnet 4.5 lists at $3 and $15.

Is Claude Sonnet 4.5 or DeepSeek V4 Flash better for coding?

They score almost the same on coding (47.3 vs 47.9); test both on your own repository before choosing.

Which has the bigger context window?

DeepSeek V4 Flash does, with 1M tokens against 200K.

How many benchmarks do Claude Sonnet 4.5 and DeepSeek V4 Flash share?

39 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and DeepSeek V4 Flash has 41.

Related comparisons

Go deeper