Model comparison

Claude Sonnet 4.5 vs DeepSeek-R1

Claude Sonnet 4.5 is the stronger model overall, scoring 44.1 to 42.3 on the Noometry Index. DeepSeek-R1 costs 6.6× less per token, which makes it the better buy when Claude Sonnet 4.5's lead doesn't matter for your workload.

Last verified October 11, 2026. 40 shared benchmarks.

Claude Sonnet 4.5 Anthropic

44.1

Rank #81 Confirmed

DeepSeek-R1

DeepSeek

42.3

Rank #115 Confirmed

Summary

They share 40 benchmarks with published results for both. Claude Sonnet 4.5 scores higher in 7 categories and DeepSeek-R1 in 2 categories; 7 gaps are clear of the uncertainty.
The widest gap is in math, where DeepSeek-R1 leads 43.8 to 32.3.
The biggest single-benchmark swing is ARC-AGI-1: 63.7% for Claude Sonnet 4.5 and 21.2% for DeepSeek-R1.
DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.5.
Claude Sonnet 4.5 accepts more context: 200K tokens versus 164K.

Side by side

Claude Sonnet 4.5 and DeepSeek-R1 specifications
	Claude Sonnet 4.5	DeepSeek-R1
Provider	Anthropic	DeepSeek
Noometry Index	44.1	42.3
Released	2025-09-29	2025-01-20
Weights	Proprietary	Proprietary
Context window	200K	164K
Max output	64K	64K
Input $ / M tokens	$3	$0.50
Output $ / M tokens	$15	$2.15
Results tracked	73	52

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

Category by category

Coding Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 47.3 (#61), DeepSeek-R1: 46.3 (#68)

Coding benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
SciCode	44.7%	35.7%
WeirdML	47.7%	41.6%
LMArena Coding	1489	1427
ALE-Bench	796.15	804.12
AlgoTune	1.52	1.7
SWE-bench Verified	71.3%	—
SWE-bench Verified (bash only)	71.4%	—
Aider Polyglot	—	71.4%
LMArena WebDev	1393	—
SWE-bench Multilingual	67%	—
GSO	14.7%	—
LiveBench Coding	—	66.7%

Agentic & Tool Use Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 38.3 (#32), DeepSeek-R1: 30.7 (#75)

Agentic & Tool Use benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
DeepResearch Bench	52.6%	35.1%
METR Time Horizons	67.4%	53.8%
Terminal-Bench	46.5%	—
Berkeley Function Calling Leaderboard	73.2%	—
GDPval	42.5%	—
Remote Labor Index	2.1%	—
τ²-bench Airline	72%	—
τ²-bench Banking	25.3%	—
τ²-bench Retail	72.4%	—
τ²-bench Telecom	84.9%	—
Cybench	60%	—
OSWorld	62.9%	—
BALROG	—	34.9%
LMArena Search	1159	—
Vending-Bench 2	3,839	—

Reasoning Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 26.9 (#125), DeepSeek-R1: 18.6 (#278)

Reasoning benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
ARC-AGI-2	13.6%	1.3%
SimpleBench	54.3%	40.8%
Kagi LLM Benchmark	57.9%	69.4%
ARC-AGI-1	63.7%	21.2%
CritPt	1.1%	1.1%
LMArena Hard Prompts	1462	1416
Epoch Capabilities Index	146.84	141.29
ForecastBench	61.9	60
NYT Connections (extended)	37.3%	—
Chess Puzzles	12%	—
EnigmaEval	6%	—
EBR-Bench	2.4%	—
LiveBench Reasoning	—	83.2%
Mystery Game Puzzles	17%	—
DTBench	83.2%	—
LiveBench Data Analysis	—	69.8%
LMCA	38.8%	—
LiveBench	—	71.6%

Math DeepSeek-R1 leads

Claude Sonnet 4.5: 32.3 (#216), DeepSeek-R1: 43.8 (#79)

Math benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
OTIS Mock AIME 2024-2025	77.8%	66.4%
Omni-MATH	55.3%	42.4%
LMArena Math	1449	1400
MATH Level 5	97.7%	96.6%
FrontierMath (Tiers 1-3)	23.9%	—
FrontierMath Tier 4	2.4%	—
ProofBench	19%	—
LiveBench Math	—	80.7%
FrontierMath (Feb 2025 set)	15.2%	—
FrontierMath Tier 4 (v1)	4.2%	—

Knowledge Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 48.4 (#76), DeepSeek-R1: 44.5 (#87)

Knowledge benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
GPQA Diamond	82.3%	76.3%
MMLU-Pro	86.9%	79.3%
Vectara Hallucination Rate	12%	11.3%
GPQA (HELM)	68.6%	66.6%
LMArena Expert	1482	1394
Humanity's Last Exam	13.7%	—
SimpleQA Verified	30.7%	—
Confabulations	—	12.7%

Multimodal Not comparable

Claude Sonnet 4.5: 34.8 (#89), DeepSeek-R1: —

Multimodal benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
VPCT	39.8%	—
LMArena Document	1450	—

Multilingual Too close to call

Claude Sonnet 4.5: 53.4 (#69), DeepSeek-R1: 52.4 (#85)

Multilingual benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
LMArena Non-English	1425	1412
LMArena Chinese	1459	1442
LMArena French	1458	1417
LMArena German	1427	1404
LMArena Japanese	1390	1391
LMArena Korean	1403	1360
LMArena Russian	1437	1423
LMArena Spanish	1457	1411

Instruction Following Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 75.0 (#78), DeepSeek-R1: 72.0 (#143)

Instruction Following benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
IFEval	85%	78.4%
LMArena Instruction Following	1459	1382
LiveBench Instruction Following	—	80.5%

Long Context Too close to call

Claude Sonnet 4.5: 45.2 (#46), DeepSeek-R1: 45.4 (#36)

Long Context benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
LMArena Longer Query	1476	1391
Fiction.LiveBench	—	75%

Writing & Preference Claude Sonnet 4.5 leads

Claude Sonnet 4.5: 66.5 (#34), DeepSeek-R1: 61.4 (#88)

Writing & Preference benchmarks
Benchmark	Claude Sonnet 4.5	DeepSeek-R1
LMArena Text	1439	1428
LMArena Creative Writing	1442	1405
EQ-Bench Creative Writing	1678	1500
WildBench	85.4%	82.8%
LMArena Multi-Turn	1465	1405
Short-Story Creative Writing	—	83%
LiveBench Language	—	48.5%

Frequently asked questions

Is Claude Sonnet 4.5 better than DeepSeek-R1?

Which is cheaper, Claude Sonnet 4.5 or DeepSeek-R1?

DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Claude Sonnet 4.5 lists at $3 and $15.

Is Claude Sonnet 4.5 or DeepSeek-R1 better for coding?

Claude Sonnet 4.5 scores higher on coding benchmarks: 47.3 versus 46.3 in the Noometry coding category.

Which has the bigger context window?

Claude Sonnet 4.5 does, with 200K tokens against 164K.

How many benchmarks do Claude Sonnet 4.5 and DeepSeek-R1 share?

40 benchmarks have published results for both models. Claude Sonnet 4.5 has 73 scored results on Noometry and DeepSeek-R1 has 52.