RankingEffort benchmarks

DeepSeek V4 Flash 0424 effort benchmarks

Compare measured Low, Medium, High and other settings. Every number keeps its source.

Model

DeepSeek V4 Flash 0424

49 measured results · 3 reported effort labels · 22 benchmark/harness combinations. Model evidence and sources →

Capability score evidence

DeepSeek V4 Flash 0424 · Max: 25.7 overall comparison score. Meets the overall evidence requirements.

Which sources carry the weight?

Artificial Analysis · 100.0% direct comparison weight · independent

These shares describe the comparisons entering the fit, not fractions of the final score.

How sensitive is this score?

Remove one evaluator or family and refit the graph against the same overall panel. Values use the matching dataset audited 2026-09-12. These are sensitivity checks, not confidence intervals; ranking eligibility may also change.

Removed evidence	Recomputed score	Change
Artificial Analysis	Comparison path lost	—
ARC Prize	28.9	+3.2
Datacurve	22.7	-3.1
LiveCodeBench	25.6	-0.1
Scale	27.3	+1.5
Terminal-Bench	25.7	+0.0
Cognition	26.3	+0.5
LiveBench	26.0	+0.3
SWE-rebench	26.1	+0.4
Epoch AI	25.7	+0.0
Vals AI	25.0	-0.7
Provider reports · Anthropic	25.7	+0.0
Provider reports · Moonshot AI	25.7	-0.0
Provider reports · Meta	25.7	-0.0
Provider reports · Mistral	25.7	+0.0
All provider reports	25.6	-0.1

Every benchmark family

Removed evidence	Recomputed score	Change
arc-agi	28.9	+3.2
deepswe	22.7	-3.0
analyst-agent	24.7	-1.0
mmmu-pro	26.7	+1.0
ifbench	24.8	-0.9
itbench	26.5	+0.8
scicode	26.4	+0.7
epoch-game-puzzles	25.0	-0.7
gpqa	25.0	-0.7
simpleqa	26.4	+0.6
tau-bench	25.1	-0.6
frontiercode	26.3	+0.5
harvey	26.2	+0.5
briefcase	26.2	+0.5
hle	26.2	+0.4
swe-rebench	26.1	+0.4
mmlu-pro	26.1	+0.4
gdp-pdf	26.1	+0.4
omniscience	26.0	+0.3
aa-lcr	25.4	-0.3
multi-swe-bench	25.4	-0.3
apex-agents	26.0	+0.3
automationbench	25.5	-0.3
livebench-instructions	25.9	+0.2
livebench-language	25.9	+0.2
swe-atlas-qna	25.5	-0.2
frontiermath	25.9	+0.2
arc-agi-3	25.6	-0.1
terminal-bench	25.8	+0.1
livebench-coding	25.6	-0.1
vibe-code	25.6	-0.1
gmmlu	25.8	+0.1
livebench-reasoning	25.6	-0.1
swe-atlas-test-writing	25.6	-0.1
vals-finance-agent	25.6	-0.1
vals-legal-research	25.6	-0.1
livecodebench	25.6	-0.1
vals-code-migration	25.7	-0.1
gdpval	25.8	+0.0
enterprise-ops	25.7	-0.0
swe-atlas-refactoring	25.7	-0.0
livebench-data	25.7	+0.0
livebench-math	25.7	+0.0
vals-excel-modeling	25.7	-0.0
aime	25.7	+0.0
enigma-eval	25.7	+0.0
swe-bench-pro	25.7	+0.0
math500	25.7	-0.0
critpt	25.7	-0.0

What evidence is missing from the fit?

Across all collected settings for this model: 49 observations, 46 contributing, 0 matched without graph weight and 3 excluded. Counts do not establish rank eligibility.

3 · Benchmark family or source protocol has not been reviewed

Inspect every observation and exclusion →

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

ADeepSeek V4 Flash 0424 · Max4 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
A29.5Supported
Hard reasoning
A33.5Supported
Coding
A22.7Supported
Human pref
Unknown
Knowledge
A36.3Supported
Multimodal
Unknown
Long context
A38.7Preliminary

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Compare this configuration →Explore the model evidence profile →

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 7,424 results across 240 models · Collected 2026-09-12

Benchmark / source	High	Max	Non-reasoning
AA-AnalystAgentArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks solved on all 5 of 5 attempts · 80 tasks, 5 attempts per task · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	25% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-09-12.	1,052.09 Elo Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	833.07 Elo Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	72% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	74.33% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	41.67% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	28.5% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	27.99% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	3.43% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	7.14% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	0.29% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
EnterpriseOps-Gym-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Overall task success rate across all eight enterprise domains (oracle tool mode) · Benchmark developed by ServiceNow Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	39.57% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
GDP.pdf: All-passArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	10.4% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	10.8% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
GDPval-AA v2 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to a human baseline of 1,000 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-09-12.	1,082.61 Elo Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	1,115.64 Elo Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
GPQA DiamondArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	86.67% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	89.39% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	71.62% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
Harvey LAB-AA: Criterion Pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of atomic pass/fail rubric criteria the deliverables satisfy, graded by a single LLM judge · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	81.33% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	30.26% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	34.85% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	7.78% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
IFBenchArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	73.47% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	79.18% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	47.21% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
ITBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Combined public and private Site Reliability Engineering (SRE) tasks · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	Not reported	31.52% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
MLCR-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of tasks judged accurate, complete, and concise · Expert + compound (hard) sets · Benchmark developed by Wisedocs · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	6.67% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	15% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-09-12.	-23.08 Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	-23.9 Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	-43.97 Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	34.87% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	36.82% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	26% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	40.16% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	45.25% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
𝜏²-Bench TelecomArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	95.61% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	95.03% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	94.44% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
𝜏³-BankingArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Sierra Research · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	26.19% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	30.93% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
Terminal-Bench HardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	38.64% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	35.61% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	34.09% Reported settings DeepSeek V4 Flash (Non-reasoning); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-non-reasoning
Terminal-Bench v2.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by the Laude Institute and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	56.93% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	61.8% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-09-12.	3.03% Reported settings DeepSeek V4 Flash (Reasoning, High Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420-high	2.53% Reported settings DeepSeek V4 Flash (Reasoning, Max Effort); Model source: https://artificialanalysis.ai/models/deepseek-v4-flash-0420	Not reported

DeepSeek V4 Flash 0424 effort benchmarks