Model comparison

DeepSeek-V3.2-Exp vs Sonar

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 38.5 on the Noometry Index.

Last verified October 10, 2026. 0 shared benchmarks.

DeepSeek-V3.2-Exp

DeepSeek

44.3

Rank #78 Confirmed

Sonar

Perplexity

38.5

Rank #187 Confirmed

Summary

The widest gap is in coding, where DeepSeek-V3.2-Exp leads 46.5 to 35.7.
DeepSeek-V3.2-Exp is cheaper at $0.26 / $0.38 per million input/output tokens, against $1 / $1 for Sonar.
DeepSeek-V3.2-Exp accepts more context: 164K tokens versus 128K.
DeepSeek-V3.2-Exp has downloadable open weights; the other is API-only.

Side by side

DeepSeek-V3.2-Exp and Sonar specifications
	DeepSeek-V3.2-Exp	Sonar
Provider	DeepSeek	Perplexity
Noometry Index	44.3	38.5
Released	2025-09-29	2024-01-01
Weights	Open	Proprietary
Context window	164K	128K
Max output	66K	4K
Input $ / M tokens	$0.26	$1
Output $ / M tokens	$0.38	$1
Results tracked	49	7

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

Category by category

Coding DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 46.5 (#65), Sonar: 35.7 (#221)

Coding benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
SWE-bench Verified (bash only)	70%	—
Aider Polyglot	74.2%	—
LMArena WebDev	1362	—
SWE-bench Multilingual	59%	—
SciCode	38.9%	—
WeirdML	39.5%	—
LiveBench Coding	—	35.1%
LMArena Coding	1454	—

Agentic & Tool Use Not comparable

DeepSeek-V3.2-Exp: 32.7 (#59), Sonar: —

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
Terminal-Bench	39.6%	—
APEX-Agents	21.3%	—
Berkeley Function Calling Leaderboard	56.7%	—
TheAgentCompany	42.9%	—
Vending-Bench 2	1,034	—

Reasoning Too close to call

DeepSeek-V3.2-Exp: 22.1 (#208), Sonar: 21.1 (#227)

Reasoning benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
ARC-AGI-2	4%	—
Kagi LLM Benchmark	52.2%	—
NYT Connections (extended)	36.7%	—
ARC-AGI-1	57%	—
CritPt	2.9%	—
Chess Puzzles	14%	—
Thematic Generalization	65%	—
LiveBench Reasoning	—	46.3%
LMArena Hard Prompts	1434	—
DTBench	87.7%	—
LiveBench Data Analysis	—	37.9%
LMCA	29.1%	—
Epoch Capabilities Index	146.27	—
LiveBench	—	46.9%

Math DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 41.7 (#87), Sonar: 33.7 (#200)

Math benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
MathArena Final-Answer Competitions	57.7%	—
OTIS Mock AIME 2024-2025	87.8%	—
ProofBench	8%	—
LiveBench Math	—	41.6%
LMArena Math	1435	—
FrontierMath (Feb 2025 set)	22.1%	—
FrontierMath Tier 4 (v1)	2.1%	—

Knowledge Not comparable

DeepSeek-V3.2-Exp: 51.7 (#66), Sonar: —

Knowledge benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
GPQA Diamond	83.4%	—
Vectara Hallucination Rate	5.3%	—
LMArena Expert	1436	—

Multilingual Not comparable

DeepSeek-V3.2-Exp: 52.2 (#90), Sonar: —

Multilingual benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
LMArena Non-English	1409	—
LMArena Chinese	1461	—
LMArena French	1433	—
LMArena German	1440	—
LMArena Japanese	1374	—
LMArena Korean	1371	—
LMArena Russian	1424	—
LMArena Spanish	1440	—

Instruction Following DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 74.5 (#93), Sonar: 71.4 (#150)

Instruction Following benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
LiveBench Instruction Following	—	76.2%
LMArena Instruction Following	1413	—

Long Context Not comparable

DeepSeek-V3.2-Exp: 47.6 (#16), Sonar: —

Long Context benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
Fiction.LiveBench	83.3%	—
CL-bench	13.2%	—
CL-bench Life	9.5%	—
LMArena Longer Query	1428	—

Writing & Preference DeepSeek-V3.2-Exp leads

DeepSeek-V3.2-Exp: 62.4 (#77), Sonar: 52.6 (#167)

Writing & Preference benchmarks
Benchmark	DeepSeek-V3.2-Exp	Sonar
LMArena Text	1425	—
LMArena Creative Writing	1403	—
EQ-Bench Creative Writing	1515	—
LMArena Multi-Turn	1427	—
LiveBench Language	—	44.1%

Frequently asked questions

Is DeepSeek-V3.2-Exp better than Sonar?

DeepSeek-V3.2-Exp is the stronger model overall, scoring 44.3 to 38.5 on the Noometry Index.

Which is cheaper, DeepSeek-V3.2-Exp or Sonar?

DeepSeek-V3.2-Exp is cheaper. It lists at $0.26 per million input tokens and $0.38 per million output tokens; Sonar lists at $1 and $1.

Is DeepSeek-V3.2-Exp or Sonar better for coding?

DeepSeek-V3.2-Exp scores higher on coding benchmarks: 46.5 versus 35.7 in the Noometry coding category.

Which has the bigger context window?

DeepSeek-V3.2-Exp does, with 164K tokens against 128K.

How many benchmarks do DeepSeek-V3.2-Exp and Sonar share?

0 benchmarks have published results for both models. DeepSeek-V3.2-Exp has 49 scored results on Noometry and Sonar has 7.