Model comparison

DeepSeek-R1 vs Qwen3.5 397B-A17B

Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index.

Last verified October 10, 2026. 22 shared benchmarks.

DeepSeek-R1

DeepSeek

42.3

Rank #115 Confirmed

Qwen3.5 397B-A17B

Alibaba (Qwen)

46.0

Rank #67 Confirmed

Summary

They share 22 benchmarks with published results for both. DeepSeek-R1 scores higher in 2 categories and Qwen3.5 397B-A17B in 7 categories; 8 gaps are clear of the uncertainty.
The widest gap is in reasoning, where Qwen3.5 397B-A17B leads 34.5 to 18.6.
The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 66.4% for DeepSeek-R1 and 88.9% for Qwen3.5 397B-A17B.
DeepSeek-R1 is cheaper at $0.50 / $2.15 per million input/output tokens, against $0.60 / $3.60 for Qwen3.5 397B-A17B.
Qwen3.5 397B-A17B accepts more context: 262K tokens versus 164K.
Qwen3.5 397B-A17B has downloadable open weights; the other is API-only.

Side by side

DeepSeek-R1 and Qwen3.5 397B-A17B specifications
	DeepSeek-R1	Qwen3.5 397B-A17B
Provider	DeepSeek	Alibaba (Qwen)
Noometry Index	42.3	46.0
Released	2025-01-20	2026-02-01
Weights	Proprietary	Open
Context window	164K	262K
Max output	64K	66K
Input $ / M tokens	$0.50	$0.60
Output $ / M tokens	$2.15	$3.60
Results tracked	52	36

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

Category by category

Coding DeepSeek-R1 leads

DeepSeek-R1: 46.3 (#68), Qwen3.5 397B-A17B: 42.0 (#114)

Coding benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Coding	1427	1465
Aider Polyglot	71.4%	—
LMArena WebDev	—	1400
SciCode	35.7%	—
WeirdML	41.6%	—
LiveBench Coding	66.7%	—
ALE-Bench	804.12	—
AlgoTune	1.7	—

Agentic & Tool Use Qwen3.5 397B-A17B leads

DeepSeek-R1: 30.7 (#75), Qwen3.5 397B-A17B: 33.3 (#53)

Agentic & Tool Use benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
APEX-Agents	—	24.9%
τ²-bench Airline	—	81.5%
τ²-bench Banking	—	9.8%
τ²-bench Retail	—	84.4%
τ²-bench Telecom	—	97.8%
DeepResearch Bench	35.1%	—
BALROG	34.9%	—
METR Time Horizons	53.8%	—

Reasoning Qwen3.5 397B-A17B leads

DeepSeek-R1: 18.6 (#278), Qwen3.5 397B-A17B: 34.5 (#70)

Reasoning benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
Kagi LLM Benchmark	69.4%	73.7%
LMArena Hard Prompts	1416	1448
Epoch Capabilities Index	141.29	146.65
ARC-AGI-2	1.3%	—
SimpleBench	40.8%	—
NYT Connections (extended)	—	58.9%
ARC-AGI-1	21.2%	—
CritPt	1.1%	—
Chess Puzzles	—	13%
Thematic Generalization	—	65.1%
LiveBench Reasoning	83.2%	—
Mystery Game Puzzles	—	18%
DTBench	—	87.5%
LiveBench Data Analysis	69.8%	—
LMCA	—	37.9%
ForecastBench	60	—
LiveBench	71.6%	—

Math Qwen3.5 397B-A17B leads

DeepSeek-R1: 43.8 (#79), Qwen3.5 397B-A17B: 46.1 (#73)

Math benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
OTIS Mock AIME 2024-2025	66.4%	88.9%
LMArena Math	1400	1454
FrontierMath (Tiers 1-3)	—	31.2%
Omni-MATH	42.4%	—
LiveBench Math	80.7%	—
MATH Level 5	96.6%	—

Knowledge Qwen3.5 397B-A17B leads

DeepSeek-R1: 44.5 (#87), Qwen3.5 397B-A17B: 53.3 (#58)

Knowledge benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
GPQA Diamond	76.3%	86.4%
LMArena Expert	1394	1462
MMLU-Pro	79.3%	—
Confabulations	12.7%	—
Vectara Hallucination Rate	11.3%	—
GPQA (HELM)	66.6%	—

Multimodal Not comparable

DeepSeek-R1: —, Qwen3.5 397B-A17B: 40.7 (#44)

Multimodal benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Vision	—	1263

Multilingual Qwen3.5 397B-A17B leads

DeepSeek-R1: 52.4 (#85), Qwen3.5 397B-A17B: 53.7 (#59)

Multilingual benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Non-English	1412	1430
LMArena Chinese	1442	1500
LMArena French	1417	1461
LMArena German	1404	1447
LMArena Japanese	1391	1426
LMArena Korean	1360	1384
LMArena Russian	1423	1429
LMArena Spanish	1411	1441

Instruction Following Qwen3.5 397B-A17B leads

DeepSeek-R1: 72.0 (#143), Qwen3.5 397B-A17B: 75.0 (#77)

Instruction Following benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Instruction Following	1382	1424
LiveBench Instruction Following	80.5%	—
IFEval	78.4%	—

Long Context DeepSeek-R1 leads

DeepSeek-R1: 45.4 (#36), Qwen3.5 397B-A17B: 44.1 (#74)

Long Context benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Longer Query	1391	1442
Fiction.LiveBench	75%	—

Writing & Preference Too close to call

DeepSeek-R1: 61.4 (#88), Qwen3.5 397B-A17B: 62.3 (#79)

Writing & Preference benchmarks
Benchmark	DeepSeek-R1	Qwen3.5 397B-A17B
LMArena Text	1428	1438
LMArena Creative Writing	1405	1401
EQ-Bench Creative Writing	1500	1478
LMArena Multi-Turn	1405	1446
Short-Story Creative Writing	83%	—
WildBench	82.8%	—
LiveBench Language	48.5%	—

Frequently asked questions

Is DeepSeek-R1 better than Qwen3.5 397B-A17B?

Qwen3.5 397B-A17B is the stronger model overall, scoring 46.0 to 42.3 on the Noometry Index.

Which is cheaper, DeepSeek-R1 or Qwen3.5 397B-A17B?

DeepSeek-R1 is cheaper. It lists at $0.50 per million input tokens and $2.15 per million output tokens; Qwen3.5 397B-A17B lists at $0.60 and $3.60.

Is DeepSeek-R1 or Qwen3.5 397B-A17B better for coding?

DeepSeek-R1 scores higher on coding benchmarks: 46.3 versus 42.0 in the Noometry coding category.

Which has the bigger context window?

Qwen3.5 397B-A17B does, with 262K tokens against 164K.

How many benchmarks do DeepSeek-R1 and Qwen3.5 397B-A17B share?

22 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and Qwen3.5 397B-A17B has 36.