Atlas

Compare Models

DeepSeek V4.1 Flash vs GPT-5.6 Terra

GPT-5.6 Terra leads by 1.3 capability points (≈95% interval −0.7 to +3.3): statistically tied · P(higher) ≈ 90%

Wins 96 of 186 shared benchmarks (70 lost, 20 tied)

Capability vs. relative cost
DeepSeek V4.1 FlashGPT-5.6 TerraClosed weightsOpen weights

Each line runs through one model’s priced reasoning settings, cheapest to most expensive; grey points are other models’ priced settings in the same range. Results whose reasoning effort the source didn’t report are left out wherever a model has named effort levels. Capability and relative cost are the published estimates (relative cost is cost per task as a multiple of GPT-5 (high)); hover a setting, or use the arrow keys, to set it beside the other model’s same setting.

DeepSeek V4.1 FlashvsGPT-5.6 Terracapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins70 – 96 · 186 shared, 20 tied

Comparing matching setups only (50 benchmarks excluded) · suites counted once
← DeepSeek V4.1 FlashGPT-5.6 Terra →

DeepSeek V4.1 FlashvsGPT-5.6 Terrawins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Arena and 36 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is GPT-5.6 Terra minus DeepSeek V4.1 Flash.

IOI (Vals v2)Code40.387.6+47.3
FrontierMath Tier 4 (v2) - PrivateMath26.870.7+43.9
RiemannBenchMath24.066.4+42.4
Surface Evolver BenchAgents46.383.8+37.5
YC-BenchGames1657236636144−35.6
HLE Extended (Mercor held-out)General QA45.370.7+25.4
ChartographyMultimodal11.634.0+22.4
MMMU-Pro Extended (Mercor held-out)Multimodal25.047.4+22.4
APEX-Agents 1.1Professional Work50.571.8+21.3
APEX-SWECode51.931.1−20.8
SWE Atlas Codebase QnA Extended (Mercor held-out)Code36.716.0−20.7
TreasuryQA BenchProfessional Work37.116.7−20.4
Concept Synth INDUCTION Challenge64 Round 1General QA17.237.5+20.3
Furniture Assembly BenchmarkMultimodal34.254.2+20.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.00.0−20.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal20.00.0−20.0
ProofBench v1.1Math54.074.0+20.0
BuseyBench SVGMultimodal4.16.1+19.9
FrontierMath Tiers 1-3 (v2) - PrivateMath67.486.0+18.6
A-Fantasia (Valid Attempts with Retries)Games59.343.3+16.0
Creative Writing v3 EloChat & Writing15401856+16.0
CritPt (avg@5)Science14.330.0+15.7
GBENCH One-shot Coding GScoreGames48.063.3+15.4
GameBench 2Games68.453.5−15.0
Wiz Cyber Model Arena v3.9Cybersecurity50.264.4+14.2
Concept Synth INDUCTION Challenge100 Round 1General QA11.025.0+14.0
Eval Connections One-Shot with TrapsGames73.086.0+13.0
Design Arena Overall Code CategoriesCode13161238−12.9
SkillsBench without SkillsAgents61.748.8−12.8
MedXpertQA MM (200-question set)Health & Medicine63.575.7+12.2
CharXiv Extended (Mercor held-out)Multimodal68.981.1+12.2
GDP.pdf (Artificial Analysis, All-pass)Professional Work12.824.6+11.8
ALE-BenchCode18292154+11.5
Extended NYT Connections 940 (Quadratic v1)Games89.678.4−11.2
Quant-Arena (iterative total score)Finance & Accounting36.547.6+11.1
ARC-AGI-2General QA72.983.9+11.0
SkillsBench with SkillsAgents69.858.9−10.9
LMArena Agent Arena Net ImprovementAgents3.90.6−10.6
Vibe Code Bench v1.1Code84.774.6−10.1
Terminal-Bench Science 0.1 (Vals)Science0.010.0+10.0
Conceptual Reasoning IndexIndexes58.465.7+10.0
HealthBench Professional (Length-Adjusted)Health & Medicine47.857.7+9.9
Humanity's Last Exam DiamondGeneral QA19.128.8+9.7
Creative Writing v3 Rubric ScoreChat & Writing14.115.9+9.3
AutomationBench-AA 1.0.6Agents68.959.6−9.2
Excel Modeling BenchmarkFinance & Accounting57.266.2+9.0
Medical Long Context Reasoning (MLCR-AA)Health & Medicine22.831.7+8.9
SRE Bench — Fully SolvedCode0.89.5+8.8
Terminal-Bench 4.0 (Artificial Analysis)Code26.835.4+8.6
LLM Chess EloGames10871293+8.5
SWE Atlas - Codebase QnACode53.545.2−8.3
SUPERChemScience56.764.9+8.2
TxBench Antibody Discovery V0Science39.031.0−8.0
CadQueryEval (Corrected Geometry)Code76.068.0−8.0
Mystery Game PuzzlesGames43.035.0−8.0
LMCAGeneral QA47.055.0+8.0
APEX-AccountingFinance & Accounting48.140.3−7.8
Harvey HLAB — Criteria Pass RateLegal87.579.8−7.7
LMArena Agent Arena Code Confirmed SuccessAgents6.6-8.7−7.6
ScBench-Long V0 VerifiedScience27.334.9+7.6
GDPval-AA v2Professional Work16321479−7.5
ACCoRDGeneral QA68.174.7+6.6
BrowseComp (Mercor 130-question subset)Search79.485.8+6.4
DeepsecBench-AA (Artificial Analysis)Cybersecurity31.225.1−6.1
GDPval-AA v2.1Professional Work16001461−6.1
LMArena Agent Arena Code Net Improvement (Steering Burden)Agents2.70.1−6.0
DeepResearch Bench II (Mercor 130-task set)Search53.247.2−6.0
GDPval (Mercor 213-task rubric score)Professional Work86.780.7−6.0
AdvancedIFChat & Writing84.078.1−5.9
Finance Agent v2 - Disclosure AnalysisFinance & Accounting61.366.9+5.6
AA-Briefcase EloAgents14241330−5.4
BioSecBench-Function V0Science38.733.5−5.2
SWE-bench VerifiedCode82.977.8−5.1
SEC-Bench Pro - release-reported cohortCybersecurity62.857.7−5.1
ReactBench (39 tasks): Fixing ReactCode18.823.8+5.0
Terminal-Bench 4.0Code31.226.3−4.9
SpatialBench V0 VerifiedScience75.470.4−4.9
Terminal-Bench-Science 0.1Science9.013.8+4.8
Senior SWE-BenchCode22.727.4+4.7
SciCode (main problems)Science11.215.8+4.6
LMArena Agent Arena Work Confirmed SuccessAgents5.7-3.4−4.6
Finance Agent v2 - Earnings AnalysisFinance & Accounting72.977.3+4.4
LMArena Agent Arena Work Net Improvement (Steering Burden)Agents2.50.7−4.4
LMArena Agent Arena Chat Confirmed SuccessAgents4.9-3.5−4.2
ITBench-AAAgents46.951.0+4.1
APEX-1 (Mercor, mean score)Professional Work65.569.5+4.0
CharXiv (descriptive and reasoning combined)Multimodal90.194.1+4.0
Longform Creative WritingChat & Writing74.078.0+4.0
SWE-bench Verified Extended (Mercor held-out)Code82.778.7−4.0
SWE-bench MultilingualCode98.294.4−3.8
LMArena Agent Arena SteerabilityAgents-3.34.1+3.7
AA-LCR Extended (Mercor held-out)Search70.874.4+3.6
Epoch Capabilities Index (ECI)Indexes156161+3.6
LMArena Agent Arena Bash RecoveryAgents7.90.9−3.5
ReactBench (39 tasks): Writing ReactCode47.050.4+3.5
DTBenchGeneral QA89.993.3+3.5
Finance Agent v2 - PrecedentsFinance & Accounting28.224.8−3.4
DistressBenchSafety79.876.4−3.3
AA-LCR v1.1Search55.358.7+3.3
LMArena Agent Arena Chat Net Improvement (Steering Burden)Agents1.60.3−3.2
SciCode 1.0.1 (Artificial Analysis)Code51.955.0+3.1
LiveBench 2.0Agents78.075.0−3.1
SpatialBench-Long V0 VerifiedScience34.937.9+3.0
GeneBench-Pro Public Release SubsetScience11.08.0−3.0
BullshitBench v2 LeaderboardSafety57.054.0−3.0
Finance Agent v2 - General Quantitative AnalysisFinance & Accounting69.872.7+2.9
Tax Agent BenchFinance & Accounting62.565.2+2.7
Finance Agent v2 - Market AnalysisFinance & Accounting63.460.6−2.7
Artificial Analysis Intelligence Index v4.3Indexes39.542.3+2.7
AA-Omniscience IndexIndexes-5.30.1+2.7
MedScribeHealth & Medicine85.582.9−2.6
Artificial Analysis Intelligence Index v4.3.2Indexes39.542.1+2.6
MedXpertQA MM Extended (Mercor held-out)Health & Medicine48.451.0+2.6
Terminal-Bench 2.1Code90.688.0−2.6
VoxelBench Text RatingGames18611894+2.4
MedCodeHealth & Medicine41.243.4+2.2
EpiBenchScience42.139.9−2.2
Code MigrationCode45.647.8+2.2
LMArena Agent Arena Chat Bash ReliabilityAgents-1.03.0+2.0
ARC-AGI-1 Semi-PrivateGeneral QA94.596.5+2.0
Terminal-Bench 2.1 Extended (Mercor held-out)Code38.436.4−2.0
Public Benefits Bench v1.1Professional Work64.362.4−1.9
Financial Coaching BenchFinance & Accounting72.871.0−1.9
LegalBenchLegal83.385.1+1.8
Political Science LLM BenchmarkProfessional Work66.568.3+1.8
Vals Index v2Indexes57.959.6+1.7
Finance Agent v2 - All PassFinance & Accounting41.142.8+1.7
AIME 2025Math98.3100.0+1.7
SAGE v1.1Math85.487.0+1.7
CyberBench v1.1 OverallCybersecurity73.772.1−1.6
OTIS Mock AIME 2024-2025Math98.399.7+1.4
Artificial Analysis Progress Index v1Indexes320324+1.3
LMArena Text Arena EloChat & Writing14751468−1.2
AML BenchFinance & Accounting63.462.3−1.1
MMMU ProMultimodal65.666.7+1.1
Chess Puzzles (task version 2.0)Games48.047.0−1.0
τ³-Banking (Mercor run, bm25_grep retrieval)Agents37.838.8+1.0
Finance Agent v2 - ComparablesFinance & Accounting46.147.0+1.0
LMArena Agent Arena Code Steering BurdenAgents2.34.2+1.0
Humanity's Last ExamGeneral QA36.835.9−0.9
Vision Arena OverallMultimodal12631266+0.9
SAGEMath47.947.0−0.9
DeepSWE v1.1Code65.664.8−0.8
SpeciEval (Refusal-Aware Retries, Historical)Safety97.498.1+0.7
SpeciEval (Pooled English Runs)Safety97.498.1+0.7
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science5.25.8+0.6
TxBench Oligo Discovery V0Science33.933.3−0.6
LMArena Agent Arena Chat Steering BurdenAgents-0.50.6+0.5
LMArena Agent Arena Code Praise vs ComplaintAgents3.04.0+0.5
LMArena Agent Arena Work Bash ReliabilityAgents0.11.1+0.5
LMArena Agent Arena Work Steering BurdenAgents1.52.5+0.5
Finance Agent v2 - AdjustmentsFinance & Accounting49.249.7+0.5
LMArena Agent Arena Code Bash ReliabilityAgents-0.20.7+0.4
Finance Agent v2 - General Qualitative AnalysisFinance & Accounting67.467.8+0.4
LMArena Agent Arena Chat Praise vs ComplaintAgents3.12.4−0.3
LMArena Agent Arena Work Praise vs ComplaintAgents3.73.2−0.2
MMLU ProGeneral QA88.288.1−0.1
LMArena Agent Arena Chat Tool HallucinationAgents0.30.5+0.1
ARC-AGI-3 (Semi-Private)General QA0.90.8−0.1
Finance Agent v2 - Financial ModelingFinance & Accounting23.123.00.0
LMArena Agent Arena Code Tool HallucinationAgents0.50.50.0
LMArena Agent Arena Work Tool HallucinationAgents0.50.50.0
Harvey's Legal Agent Benchmark - Antitrust CompetitionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Arbitration International Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Banking FinanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Bankruptcy RestructuringLegal0.00.00.0
Harvey's Legal Agent Benchmark - Capital MarketsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.00.00.0
Harvey's Legal Agent Benchmark - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Environmental ESGLegal0.00.00.0
Harvey's Legal Agent Benchmark - Funds Asset ManagementLegal0.00.00.0
Harvey's Legal Agent Benchmark - ImmigrationLegal0.00.00.0
Harvey's Legal Agent Benchmark - InsuranceLegal0.00.00.0
Harvey's Legal Agent Benchmark - Intellectual PropertyLegal0.00.00.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Litigation Dispute ResolutionLegal0.00.00.0
Harvey's Legal Agent Benchmark - Real EstateLegal0.00.00.0
Harvey's Legal Agent Benchmark - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0
Legal Research BenchLegal41.341.30.0
ProgramBenchCode0.50.50.0

DeepSeek V4.1 FlashvsGPT-5.6 Terranormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Arena and 36 othersVillage LLM Atlasatlas.village.dev

Win counts use each benchmark’s declared score direction. Represented suite slices and the derived Capability score are excluded from headline totals. Missing or incompatible setups are excluded, never counted as losses.
Loading Atlas data…