DeepSeek-V4-Flash

DeepSeek-V4-Flash places in the top three on 3 benchmarks.

DeepSeek Released 24 Apr 2026

Benchmarks with a score

Ranked first, out of five or more models

In the top three

DeepSeek Apr 2026

Open weights

146.2 #83 of 271

Epoch Capabilities Index

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.

Benchmark	Score	Rank	Source
MMLU-Pro	86.2%	#17 of 151	Model card
Chinese SimpleQA	78.9%	#3 of 13	Model card
Extended NYT Connections	89.6%	#23 of 97	Official leaderboard
LMArena (Elo)	1,439	#79 of 333	Official leaderboard
EQ-Bench 3	1,504	#22 of 78	Official leaderboard
Wer wird Millionär? LLM Benchmark	61867.0	#24 of 76	Official leaderboard
MRCR v2	78.7%	#10 of 27	Model card
Longform Creative Writing (EQ-Bench)	66.0	#53 of 119	Official leaderboard
Humanity's Last Exam	34.8%	#68 of 147	Model card
Creative Writing v3	69.8	#57 of 120	Official leaderboard
SEA-HELM	69.8	#32 of 62	Official leaderboard
SimpleQA Verified	34.1%	#13 of 22	Model card
Phare	63.7%	#40 of 62	Official leaderboard
BullshitBench	16.4%	#80 of 107	Official leaderboard
BullshitBench V2	18%	#92 of 122	Official leaderboard
Judgemark v4	36.8	#31 of 40	Official leaderboard
CorpusQA 1M	60.5%	#2 of 2	Model card

Benchmark	Score	Rank	Source
GPQA Diamond	88.1%	#48 of 284	Model card
IMO-AnswerBench	88.4%	#8 of 37	Model card
MathArena (Apex, IMO, USAMO)	27.1%	#11 of 47	Official leaderboard
SimpleBench	61.1%	#34 of 100	Official leaderboard
HMMT 2026	93.9%	#11 of 31	Official leaderboard
DTBench	77.3%	#66 of 178	Official leaderboard
CritPt	7.14%	#66 of 160	Epoch AI
PutnamBench	81%	#3 of 7	Model card
xbench-ScienceQA	71%	#14 of 29	Official leaderboard
AIME 2026	95.8%	#17 of 31	Official leaderboard
Verifying the Verifiers (Project APE CRED benchmark)	80.5%	#13 of 23	Official leaderboard

Benchmark	Score	Rank	Source
LiveCodeBench v6	91.6%	#5 of 76	Model card
Codeforces rating	3,052	#4 of 41	Model card
StructEval	78.3%	#4 of 28	Official leaderboard
SWE-bench Verified	79%	#22 of 150	Model card
Code Arena	1,582	#35 of 124	Official leaderboard
SWE-bench Multilingual	73.3%	#34 of 70	Model card
WeirdML	45.6%	#60 of 112	Official leaderboard
EvoCode-Bench v2	12.2%	#7 of 13	Official leaderboard
WebDev Arena	1,431	#62 of 105	Epoch AI
ALE-Bench	678	#83 of 121	Epoch AI
Tencent WorkBuddy Bench	74.9%	#7 of 10	Official leaderboard
DeepSWE	53.3%	#20 of 28	Official leaderboard
SWE-Bench Pro	52.6%	#57 of 77	Model card
Monthly-SWEBench	45%	#9 of 12	Official leaderboard
SWE Refactor Bench	7.00%	#6 of 8	Official leaderboard
ReviewBench	64%	#3 of 4	Official leaderboard
EvoCode-Bench	9.40%	#10 of 13	Official leaderboard
AndroidBench	54.7%	#26 of 32	Official leaderboard

Benchmark	Score	Rank	Source
Toolathlon	48.2%	#10 of 50	Official leaderboard
PinchBench 2.0	81.7%	#14 of 57	Official leaderboard
LisanBench	2789.0	#36 of 133	Official leaderboard
GAIA	69.1%	#18 of 57	Official leaderboard
Long-Horizon-Terminal-Bench (LHTB)	45.5%	#8 of 25	Official leaderboard
Terminal-Bench 2.0	56.9%	#28 of 61	Model card
Long-Horizon-Terminal-Bench 2.0	52.4%	#7 of 15	Official leaderboard
BrowseComp	73.2%	#42 of 77	Model card
Claw-Eval Pass^3	57.8%	#14 of 23	Official leaderboard
CivBench	1,477	#12 of 18	Official leaderboard
MCP Atlas	69%	#27 of 40	Model card
SkillsBench	44.7%	#15 of 21	Official leaderboard
Claw-Eval	74.2%	#17 of 23	Official leaderboard
Toolathlon-Verified	50.9%	#19 of 25	Official leaderboard
LiveClawBench	60.5%	#13 of 17	Official leaderboard
ClawBench V2	2.30%	#6 of 7	Official leaderboard
PACT	1,483	#23 of 26	Official leaderboard

Benchmark	Score	Rank	Source
FORTE	70.7%	#9 of 21	Official leaderboard
CommerceAgentBench	46.7%	#8 of 13	Official leaderboard
Clipping Bench	$320	#17 of 19	Official leaderboard
GDPval-AA	1,395	#15 of 16	Model card

Benchmark	Score		Rank	Source
ScienceVision	0.27		#10 of 11	Official leaderboard

Benchmark	Score	Rank	Source
CyberGym	97.1%	#3 of 33	Official leaderboard
DeepsecBench	16.4	#16 of 28	Official leaderboard
CyberBench (Vals)	66.7%	#16 of 26	Official leaderboard

Benchmark	Score	Rank	Source
Political Science LLM Benchmark	66.5%	#12 of 29	Official leaderboard
LMCA	35.9%	#68 of 139	Epoch AI
Conceptual Reasoning Index	48.5	#73 of 149	Official leaderboard
SuperCLUE	65.6	#10 of 18	Official leaderboard
ACCoRD	57.3	#95 of 150	Official leaderboard

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

Every result in the longer documents the lab published with it. 21 of them aren't in the launch post.