DeepSeek V4.1 Flash

DeepSeek V4.1 Flash ranks first on 4 of the 153 benchmarks it has scores on.

DeepSeek Released 9 Sep 2026 Family DeepSeek-V4.1 The lab's model card

155

Benchmarks with a score

Ranked first, out of five or more models

In the top three

DeepSeek Sep 2026

DeepSeek V4.1 Flash

Open weights Mid-size

156.1 #34 of 271

Epoch Capabilities Index

MMMU-Pro 56.5% #61 /72
NL2Repo-Bench 65.4% #2 /28
LongBench v2 45.2% #24 /27

Where it ranks

Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.

Knowledge and language

Benchmark	Score	Rank	Source
Vectara Hallucination Leaderboard (2025 dataset)	4.20%	#5 of 104	Official leaderboard
Humanity's Last Exam	63.9%	#8 of 147	Model card
LMArena (Elo)	1,476	#30 of 333	Official leaderboard
AGIEval	83.4%	#4 of 35	Model card
C-Eval	92.1%	#5 of 38	Model card
Wer wird Millionär? LLM Benchmark	175000.0	#14 of 76	Official leaderboard
Longform Creative Writing (EQ-Bench)	74.0	#29 of 119	Official leaderboard
Extended NYT Connections	89.6%	#24 of 97	Official leaderboard
DROP	87.9%	#11 of 40	Model card
BullshitBench V2	57%	#36 of 122	Official leaderboard
BIG-Bench Hard	86.1%	#22 of 71	Model card
BullshitBench	41.8%	#34 of 107	Official leaderboard
HellaSwag	87.2%	#15 of 47	Model card
SimpleQA Verified	42.3%	#21 of 56	Official leaderboard
Creative Writing v3	70.3	#53 of 120	Official leaderboard
MGSM	80.2%	#30 of 52	Model card
BIG-Bench Extra Hard	27.2%	#7 of 11	Model card
MMLU-Pro	74.1%	#104 of 151	Model card
SuperGPQA	53.1%	#23 of 26	Model card
LongBench v2	45.2%	#24 of 27	Model card
AI Detection Benchmark	65.4%	#3 of 3	Official leaderboard
MultiLoKo	45.5%	#2 of 2	Model card

Math and science

Benchmark	Score	Rank	Source
AIME 2026	100%	#1 of 34	Model card
OTIS Mock AIME 2024-2025	98.3%	#27 of 202	Epoch AI
GPQA Diamond	89.8%	#45 of 219	Epoch AI
ARC-AGI	94.5%	#21 of 92	Official leaderboard
SimpleBench	66.7%	#23 of 100	Official leaderboard
ARC-AGI-2	72.9%	#22 of 95	Official leaderboard
xbench-ScienceQA	72.6%	#8 of 29	Official leaderboard
DTBench	83.1%	#52 of 178	Official leaderboard
GSM8K	93%	#26 of 87	Model card
CritPt	14.3%	#48 of 160	Epoch AI
ProofBench	54%	#29 of 83	Epoch AI
FrontierMath Tiers 1-3 v2	67.4%	#33 of 89	Epoch AI
Terminal-Bench-Science	15.7%	#7 of 17	Official leaderboard
Prophet Arena	85.4	#8 of 18	Official leaderboard
MathArena (Apex, IMO, USAMO)	65.6%	#6 of 13	Model card
BrokenArXiv	49.6%	#3 of 6	Official leaderboard
LatchBio Capabilities v1.0	42.1%	#14 of 27	Official leaderboard
ProofBench v1.1	54%	#28 of 51	Official leaderboard
MATH	61.1%	#52 of 93	Model card
FrontierMath Tier 4 v2	26.8%	#42 of 71	Epoch AI
Surface Evolver Bench	46.3%	#17 of 26	Epoch AI
MysteryMechanism	21.2%	#23 of 29	Official leaderboard
ExplorationBench	59.5%	#8 of 10	Official leaderboard
PaperBenchX	7.53%	#8 of 10	Official leaderboard
ArXivMath	47.4%	#5 of 6	Official leaderboard
ArXivLean	5.66%	#6 of 7	Official leaderboard

Coding

Benchmark	Score	Rank	Source
Codeforces rating	3,471	#1 of 41	Model card
NL2Repo-Bench	65.4%	#2 of 28	Model card
BigCodeBench	60.6%	#1 of 11	Model card
Code Arena	1,665	#15 of 124	Official leaderboard
SWE-bench-Live	71%	#2 of 13	Official leaderboard
DeepSWE	74.2%	#8 of 51	Model card
Vibe Code Bench	84.7%	#19 of 115	Official leaderboard
WebDev Arena	1,614	#18 of 105	Epoch AI
Code Migration	45.6%	#17 of 78	Official leaderboard
Image-to-WebDev Arena	1,606	#13 of 55	Official leaderboard
Code Migration v1.1	58.9%	#9 of 29	Official leaderboard
FrontierCode	43.9%	#17 of 46	Epoch AI
ScarfBench	69%	#4 of 10	Official leaderboard
ALE-Bench	1,092	#50 of 121	Epoch AI
HumanEval	79.4%	#52 of 95	Model card
OpenDesign Arena	81.2	#11 of 19	Official leaderboard
Brokk Power Ranking	7.30	#12 of 19	Official leaderboard
Senior SWE-Bench	22.7%	#15 of 23	Official leaderboard
Vibe Code Bench 1-100	16.4%	#17 of 26	Official leaderboard
FrontierCS 2	14.6%	#12 of 16	Official leaderboard
IOI v2 (Vals)	40.3%	#35 of 45	Official leaderboard
WeirdML v3	5.95%	#21 of 25	Official leaderboard
FrontierSWE	32.9%	#8 of 9	Model card
ProgramBench	20.3%	#18 of 20	Model card
ViBench 2.0	0%	#9 of 9	Official leaderboard
RSI Bench	32.1	#8 of 8	Official leaderboard

Agents and tools

Benchmark	Score	Rank	Source
Terminal-Bench 2.1	90.6%	#3 of 53	Model card
AutomationBench	54.8%	#5 of 41	Model card
LLM Chess	1,087	#20 of 124	Official leaderboard
Mystery Game Puzzles	43%	#17 of 80	Epoch AI
Chess Puzzles	48%	#21 of 83	Epoch AI
CivBench	1,537	#5 of 18	Official leaderboard
Agent Arena	0.03	#18 of 53	Official leaderboard
Terminal-Bench 3.0	30%	#2 of 5	Model card
ARC-AGI-3	3.26%	#10 of 24	Official leaderboard
Hermes Bench	70.7%	#10 of 17	Official leaderboard
GAIA	41.2%	#35 of 57	Official leaderboard
Terminal-Bench 4.0	31.2%	#19 of 23	Model card

Professional work

Benchmark	Score	Rank	Source
Vals Finance Agent v2.1	65.4%	#8 of 35	Official leaderboard
MedScribe	85.5%	#26 of 98	Official leaderboard
BlenderWorld	65.7	#4 of 13	Official leaderboard
Legal Research Bench	41.4%	#29 of 88	Official leaderboard
PolicyBench	86.3%	#19 of 50	Official leaderboard
Agents' Last Exam	31.8%	#10 of 25	Model card
LegalBench	83.3%	#56 of 138	Official leaderboard
Harvey LAB (Legal Agent Benchmark)	6.67%	#34 of 80	Official leaderboard
Vals Finance Agent v2	53.5%	#34 of 80	Official leaderboard
RCM-Bench-1	89%	#9 of 20	Official leaderboard
Public Benefits Bench	64.3%	#23 of 50	Official leaderboard
Tax Agent Bench	26.6%	#37 of 79	Official leaderboard
Vals Finance Index	57.1%	#17 of 35	Official leaderboard
Excel Modeling Benchmark	57.2%	#40 of 76	Official leaderboard
Physical-Engineering's Last Exam	0.23	#10 of 18	Official leaderboard
MedCode	41.2%	#54 of 96	Official leaderboard
Thrive Holdings helpdesk ticket resolution benchmark	48.1%	#9 of 15	Official leaderboard
Immigration Bench	65.6%	#16 of 24	Official leaderboard
Excel Modeling Benchmark v1.1	67.8%	#25 of 37	Official leaderboard
OR-License	69.7%	#7 of 10	Official leaderboard
Legora BAR	100.0	#12 of 17	Official leaderboard
APEX-Agents	39.5%	#44 of 56	Epoch AI
RWE-bench	0%	#9 of 11	Official leaderboard
TaxCalcBench v2	4.00%	#22 of 26	Official leaderboard
Holdings-IndividualTaxBench	69.9%	#11 of 13	Official leaderboard
FoodTruck Bench	$50,179	#9 of 10	Official leaderboard
Quant-Arena	36.5	#9 of 9	Official leaderboard

Vision, audio and robotics

Benchmark	Score	Rank	Source
ZeroBench	49%	#4 of 25	Model card
DocVQA	95.6%	#6 of 33	Model card
BabyVision	89.6%	#5 of 20	Model card
Vision Arena	1,261	#39 of 136	Official leaderboard
VoxelBench	1,858	#18 of 56	Official leaderboard
Chartography	78.9%	#4 of 12	Model card
VibeVideoBench	3.91	#3 of 9	Official leaderboard
GDP.pdf	19.8%	#23 of 39	Epoch AI
Furniture Assembly Benchmark	34.2%	#23 of 35	Epoch AI
RoboColiseum	55.1%	#6 of 9	Official leaderboard
MMMU-Pro	56.5%	#61 of 72	Model card
CVBench	77.9%	#11 of 11	Model card
DataCenterBench Tier 1	11.7%	#9 of 9	Official leaderboard
Embodied First Exam	2.63%	#7 of 7	Official leaderboard

Safety and security

Benchmark	Score	Rank	Source
CyberGym	99.4%	#1 of 33	Official leaderboard
DistressBench	79.8%	#5 of 42	Official leaderboard
BioSecBench-Surveillance	51.5%	#5 of 27	Official leaderboard
CyberBench (Vals)	78.8%	#5 of 26	Official leaderboard
CyberBench v1.1 (Vals)	73.7%	#9 of 43	Official leaderboard
EpiBench	25.5%	#10 of 27	Official leaderboard
ExploitGym	23.7%	#7 of 18	Official leaderboard
BioSecBench-Function	40.8%	#11 of 27	Official leaderboard
TxBench-Antibody-Discovery	39.7%	#14 of 28	Official leaderboard
MetagenomicsBench	38%	#14 of 26	Official leaderboard
TxBench-Oligo-Discovery	34.8%	#16 of 28	Official leaderboard
scBench-Long	14.8%	#17 of 27	Official leaderboard
SpatialBench-Long	26.2%	#17 of 27	Official leaderboard
BioSecBench-Refusal	11.3%	#14 of 22	Official leaderboard
TxBench-Preclinical-Pharmacology	46.7%	#18 of 28	Official leaderboard
SEC-bench Pro	62.8%	#7 of 10	Model card
OncologyBench	14.7%	#22 of 30	Official leaderboard
VariantBench	31.1%	#20 of 27	Official leaderboard
CWE-bench	55%	#9 of 12	Official leaderboard
SRE-Bench	0.76%	#27 of 33	Official leaderboard

Other

Benchmark	Score	Rank	Source
SuperCLUE	71.8	#2 of 18	Official leaderboard
LMCA	47%	#33 of 139	Epoch AI
Conceptual Reasoning Index	58.4	#43 of 149	Official leaderboard
SAGE (Vals)	47.9%	#29 of 81	Official leaderboard
ACCoRD	68.1	#56 of 150	Official leaderboard
SAGE v1.1 (Vals)	85.4%	#16 of 28	Official leaderboard
Hermes Index	36.9	#11 of 17	Official leaderboard
Vals Science Index	29.6%	#18 of 23	Official leaderboard

DeepSeek's models over time

DeepSeek's models in color, against every other model on the Epoch Capabilities Index, by release date.

Its launch post

The benchmarks the lab chose to headline, with the scores it gave.

DeepSeek-V4.1-Flash model card 10 Sep 2026

GPQA Diamond (Pass@1) 90.9
Best of HLE (Pass@1) / HLE w/ tools (Pass@1) 63.9
Codeforces (Rating) 3471
MathArena Apex (Pass@1) 65.6
Terminal-Bench 2.1 (Pass@1) 90.6
Terminal-Bench 3.0 (Pass@1) 30
Terminal-Bench 4.0 (Pass@1) 31.2
DeepSWE v1.1 (Resolved) 74.2
ProgramBench (Almost@1) 20.3
NL2Repo-Bench (Score) 64
CyberGym (Pass@1) 88.1
SEC-Bench Pro (Pass@1) 62.8
ExploitGym (Pass@1) 15.3
AutomationBench (Pass@1) 54.8
Agent's Last Exam (Pass@1) 31.8
Chartography w/ tools (Pass@1) 78.9
BabyVision w/ tools (Pass@1) 89.6
ZeroBench-main w/ tools (Pass@5) 49
AGIEval (EM) 83.4
MMLU-Pro (EM) 74.1
C-Eval (EM) 92.1
MultiLoKo (LLM-Judge) 45.5
SimpleQA-Verified (EM) 42.3
SuperGPQA (EM) 53.1
BBH (EM) 86.1
BBEH (EM) 27.2
DROP (F1) 87.9
HellaSwag (EM) 87.2
BigCodeBench (Pass@1) 60.6
HumanEval (Pass@1) 79.4
GSM8K (EM) 93
MATH (EM) 61.1
MGSM (EM) 80.2
LongBench-V2 (EM) 45.2
MMMU-Pro (EM) 56.5
CVBench (EM) 77.9
DocVQA (LLM-Judge) 95.6

Its system card

Every result in the longer documents the lab published with it. 4 of them aren't in the launch post.