DeepSeek V4 Flash 0420

Across the task benchmarks we track, DeepSeek V4 Flash 0420’s strongest relative results are in coding, terminal use, and work automations. On SWE-bench Verified, it scores 79%, ranking #4 among 7 models with reported results in our dataset.

Updated: Nov 5, 2026Data as of: Nov 2, 2026

Compare this model

Strongest results

Strongest results for DeepSeek V4 Flash 0420 in tracked evaluation cohorts
Benchmark	Measures	Score
IFBench ↗58 constraint types	Precise instruction following	79.18%#2 of 17
SWE-bench ↗Verified / published open-model reports	Coding	79%#4 of 7
Humanity's Last Exam ↗Stable HLE — text-only subset	Expert reasoning	34.85%#43 of 68
AA-Briefcase ↗1.1 / combined Elo	Business deliverables	1,041.27 Elo#39 of 57

Model details

Open source / open weights13 benchmarks

Specifications for DeepSeek V4 Flash 0420
Model	DeepSeek V4 Flash 0420
Agent Index v1.2	Insufficient coverage (3/7) · Preview
Creator	DeepSeek (opens in a new tab)
Availability	Open source / open weights
Context size	1,000,000 tokens (opens in a new tab)
Max output	Not reported
Knowledge cutoff	Not reported
License	MIT · Model weights ↗ (opens in a new tab)

Pricing and performance

Per provider and reported configuration. Prices are USD per 1M tokens.

No provider pricing or performance reports have been added for this model yet.

Benchmark performance

DeepSeek V4 Flash 0420 in color, alongside leading models. Latest stored reports, best effort per report date.

Reasoning

198-question Diamond subset; independent AA evaluation

GPQA Diamond ↗

01
GPT-6 Astra96.3%
02
Gemini 3.8 Flash95.3%
03
Gemini 3.1 Pro94.1%
03
GPT-5.6 Sol94.1%
03
Muse Spark 1.394.1%
24
DeepSeek V4 Flash 0420This model89.4%

Coding

Verified split; published reports

SWE-bench Verified ↗

Terminal Use

89 tasks; Terminus 2; e2b; three repeats

Terminal-Bench 2.1 ↗

01
Kimi K385%
02
GLM-5.3-Flash84.3%
03
GLM-5.383.9%
04
DeepSeek V4 Pro 081378.7%
05
GLM-5.277.9%
09
DeepSeek V4 Flash 0420This model61.8%

Web Research

Published reports; source settings retained

BrowseComp ↗

01
Kimi K3.1 Pro 92.1%
02
Kimi K3 91.2%
03
MiniMax M3 83.5%
04
DeepSeek V4 Pro 0424 83.4%
05
Kimi K2.6 83.2%
07
DeepSeek V4 Flash 0420This model73.2%

Work Automations

Private 657 tasks; 50 turns; objective completion

AutomationBench-AA ↗

01
Gemini 4 Carbon77.8%
02
Gemini 4 Argon77.5%
03
Kimi K3.1 Pro73.3%
04
MiniMax M3.1 Flash72.52%
05
Claude Sonnet 5.571.8%
39
DeepSeek V4 Flash 0420This model28.5%

Knowledge Work

220 tasks; Stirrup agent; pairwise Elo

GDPval-AA v2.1 ↗

01
Claude Opus 5.51,866 Elo
02
Kimi K3.1 Pro1,858 Elo
03
Claude Sonnet 5.51,838 Elo
04
Claude Fable 5.11,763 Elo
05
DeepSeek V4.1 Flash 10171,746 Elo
32
DeepSeek V4 Flash 0420This model1,046 Elo

Reported benchmarks

All published results in our dataset, with versions, evaluation settings, and sources.

All published benchmark reports for DeepSeek V4 Flash 0420
Benchmark	Score	Evaluation setup	Source
AA-Briefcase1.1 / combined EloCombined Elo · Higher is better	1,041.27 Elo	selection: best reported configuration per release · DeepSeek V4 Flash 0420 (High) · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
AutomationBench-AA1.0.6 / AA private 657Objectives completed without guardrail violations · Higher is better	27.986%	DeepSeek V4 Flash 0420 (Max); Private 657 tasks; 50 turns; objective completion	Artificial Analysis (opens in a new tab)independent
AutomationBench-AA1.0.6 / AA private 657Objectives completed without guardrail violations · Higher is better	28.504%	DeepSeek V4 Flash 0420 (High); Private 657 tasks; 50 turns; objective completion	Artificial Analysis (opens in a new tab)independent
BrowseCompPublished open-model reportsAccuracy · Higher is better	73.2%	harness comparison: provider-reported; settings differ across models · Think Max; provider browsing agent; discard-all context management; original V4 release	DeepSeek ↗ (opens in a new tab)provider reported
GDP.pdf100 tasks / All-pass / five attemptsAccuracy · Higher is better	10.8%	selection: best reported configuration per release · DeepSeek V4 Flash 0420 (Max) · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
GDPval-AA2.1Elo · Higher is better	1,045.62 Elo	DeepSeek V4 Flash 0420 (Max); 220 tasks; Stirrup agent; pairwise Elo	Artificial Analysis (opens in a new tab)independent
GDPval-AA2.1Elo · Higher is better	1,009.33 Elo	DeepSeek V4 Flash 0420 (High); 220 tasks; Stirrup agent; pairwise Elo	Artificial Analysis (opens in a new tab)independent
GPQADiamond / Artificial AnalysisAccuracy · Higher is better	89.394%	DeepSeek V4 Flash 0420 (Max); 198-question Diamond subset; independent AA evaluation	Artificial Analysis (opens in a new tab)independent
GPQADiamond / Artificial AnalysisAccuracy · Higher is better	71.616%	DeepSeek V4 Flash 0420 (Non-reasoning); 198-question Diamond subset; independent AA evaluation	Artificial Analysis (opens in a new tab)independent
GPQADiamond / Artificial AnalysisAccuracy · Higher is better	86.667%	DeepSeek V4 Flash 0420 (High); 198-question Diamond subset; independent AA evaluation	Artificial Analysis (opens in a new tab)independent
Humanity's Last ExamHLE — source does not specify subsetAccuracy · Higher is better	45.1%	Original V4 Think Max; HLE with tools; source does not explicitly identify subset · Tools enabled	DeepSeek ↗ (opens in a new tab)provider reported
Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better	34.847%	DeepSeek V4 Flash 0420 (Max); original HLE text-only 2,158 questions; no tools; not HLE Diamond · No tools	Artificial Analysis (opens in a new tab)independent
Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better	7.785%	DeepSeek V4 Flash 0420 (Non-reasoning); original HLE text-only 2,158 questions; no tools; not HLE Diamond · No tools	Artificial Analysis (opens in a new tab)independent
Humanity's Last ExamStable HLE — text-only subsetAccuracy · Higher is better	30.259%	DeepSeek V4 Flash 0420 (High); original HLE text-only 2,158 questions; no tools; not HLE Diamond · No tools	Artificial Analysis (opens in a new tab)independent
IFBench58 constraint typesAccuracy · Higher is better	79.184%	selection: best reported configuration per release · DeepSeek V4 Flash 0420 (Max) · tools: benchmark-defined	Artificial Analysis (opens in a new tab)independent
SWE-benchVerified / published open-model reportsResolved · Higher is better	79%	harness comparison: provider-reported; settings differ across models · DeepSeek original V4 release; Think Max; published SWE Verified resolved score; original release, not 0813	DeepSeek ↗ (opens in a new tab)provider reported
Terminal-Bench2.0 / published open-model reportsTask success · Higher is better	56.9%	harness comparison: provider-reported; settings differ across models · Published Terminal-Bench 2.0 score; not 2.1	DeepSeek ↗ (opens in a new tab)provider reported
Terminal-Bench2.1 / Artificial AnalysisPass@1 · Higher is better	61.798%	DeepSeek V4 Flash 0420 (Max); 89 tasks; Terminus 2; e2b; three repeats	Artificial Analysis (opens in a new tab)independent
Terminal-Bench2.1 / Artificial AnalysisPass@1 · Higher is better	56.929%	DeepSeek V4 Flash 0420 (High); 89 tasks; Terminus 2; e2b; three repeats	Artificial Analysis (opens in a new tab)independent
Terminal-Bench4.0 / mini-swe-agentTask success · Higher is better	3.03%	harness: mini-swe-agent · repeats: 3 · selection: best reported configuration per release · DeepSeek V4 Flash 0420 (High)	Artificial Analysis (opens in a new tab)independent