LetTokenBurnBenchHub

Model

DeepSeek V4 Flash

DeepSeek · open weights

Knowledge & Reasoning

Knowledge & Reasoning

Benchmark	Score	Metric	Harness	Source	Confidence	Link
CritPt	7.1	accuracy	N/A	Independent	low	epoch.ai
HouseholdBench	1.0	relmae	N/A	Independent	medium	arxiv.org
MechHypoBench	49.2	mca_(%)	codex	Independent	medium	arxiv.org

Math

Benchmark	Score	Metric	Harness	Source	Confidence	Link
EfficientOpt	85.3	solver_time_ratio	N/A	Independent	medium	arxiv.org

Coding

Benchmark	Score	Metric	Harness	Source	Confidence	Link
ArchitectureIQ	70.6	accuracy	architectureiq	Independent	medium	arxiv.org
ArrivalBench	52.4	failure_rate	N/A	Independent	medium	arxiv.org
HealBench	20.0	resolved_rate	codex	Independent	medium	arxiv.org
SciCode	44.9	score	N/A	Independent	medium	artificialanalysis.ai
SWE-CC	93.4	resolve_rate	mini-swe-agent	Independent	medium	arxiv.org
TestSuiteBench	28.6	score	N/A	Independent	medium	arxiv.org

Agentic / Tool Use

Agentic / Tool Use

Benchmark	Score	Metric	Harness	Source	Confidence	Link
AgBench	0.0	api_cost	<null>	Independent	medium	arxiv.org
AutoDataBench	55.2	performance_of_models_trained_on_optimized_data_vs_baseline_prediction_accuracy_comparison	N/A	Independent	medium	arxiv.org

Multimodal

Benchmark	Score	Metric	Harness	Source	Confidence	Link
NextAct	0.3	embedding_score	N/A	Independent	medium	arxiv.org

Safety & Alignment

Safety & Alignment

Benchmark	Score	Metric	Harness	Source	Confidence	Link
AgentDoxx	90.2	accuracy	N/A	Independent	medium	arxiv.org
APEX-Harm	47.3	success_rate	N/A	Independent	medium	arxiv.org
CRIME	37.9	success_rate	N/A	Independent	medium	arxiv.org
DecepEval	78.4	deception_rate	N/A	Independent	medium	arxiv.org
DRYLAB-Bench	0.4	floor-included_brs_s2	N/A	Independent	medium	arxiv.org
Mental Health Stigma Benchmark	0.6	f1	N/A	Independent	medium	arxiv.org

Multilingual

Benchmark	Score	Metric	Harness	Source	Confidence	Link
TRuST	48.2	recall	t-search	Independent	medium	arxiv.org

Reproduced by the community