Atlas

Compare Models

Claude Opus 5.5 vs DeepSeek V4.1 Flash 1017

Claude Opus 5.5 leads by 6.8 capability points (≈95% interval +4.2 to +9.4): a clear lead over DeepSeek V4.1 Flash 1017 · P(higher) ≈ 100%

Wins 113 of 151 shared benchmarks (20 lost, 18 tied)

Capability vs. relative cost
Claude Opus 5.5DeepSeek V4.1 Flash 1017Closed 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.

Claude Opus 5.5vsDeepSeek V4.1 Flash 1017capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins113 – 20 · 151 shared, 18 tied

Comparing matching setups only (12 benchmarks excluded) · suites counted once
← Claude Opus 5.5DeepSeek V4.1 Flash 1017 →

Claude Opus 5.5vsDeepSeek V4.1 Flash 1017wins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, SecureBio and 48 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is DeepSeek V4.1 Flash 1017 minus Claude Opus 5.5.

HardGeoBench Avg DistanceMultimodal6693201−78.1
AA-PersonalityBench Imagination indexChat & Writing94.337.2−57.7
AA-PersonalityBench Warmth indexChat & Writing84.827.7−57.7
HardGeoBench <=100miMultimodal61.37.5−53.8
HardGeoBench Country AccuracyMultimodal67.515.0−52.5
FrontierMath Tier 4 (v2) - PrivateMath95.046.3−48.7
HoneyBench v0.2 (hack rate)Safety31.676.4−44.8
Concept Synth INDUCTION Challenge100 Round 1General QA65.021.0−44.0
DAYJOB: Healthcare (October 2026 rescoring)Health & Medicine48.36.0−42.3
ArXivMath 08/2026 (Native Harnesses)Math86.043.9−42.1
RiemannBench (November 2026 rescoring)Math80.038.4−41.6
ArXivMath 10/2026 (Native Harnesses)Math79.838.3−41.5
Concept Synth INDUCTION Challenge64 Round 1General QA71.932.8−39.1
HardGeoBench <=10miMultimodal42.53.8−38.7
Terminal-Bench-Science 0.1Science49.010.5−38.6
AA-PersonalityBench Expressiveness indexChat & Writing90.552.4−38.5
ArXivMath 09/2026 (Native Harnesses)Math82.145.3−36.8
FrontierML ($200 wallet)Agents58.823.5−35.4
DAYJOB: Data ScienceProfessional Work37.33.3−34.0
Terminal-Bench 4.0Code64.831.0−33.8
Humanity's Last Exam DiamondGeneral QA54.620.8−33.8
EmbodiedLab (T0 success)Agents43.310.0−33.3
LMArena Agent Arena Net Improvement (Steering Burden)Agents19.52.0−33.1
CritPt v2 (agentic, avg@5)Science81.750.6−31.1
BrokenArXiv 08/2026 (Native Harnesses)Math77.446.7−30.6
DAYJOB: Finance (October 2026 rescoring)Finance & Accounting42.512.5−30.0
TCSExam LiteMath72.342.4−29.9
SimpleQA Verified (Epoch anti-abstention)General QA72.242.5−29.7
FrontierCode 2.0 - Pass RateCode51.823.0−28.8
HANDBOOK.md Agents v1.1Professional Work52.324.0−28.3
BrokenArXiv 10/2026 (Native Harnesses)Math77.849.7−28.1
Terminal-Bench 4.0 (Artificial Analysis)Code56.628.8−27.8
FrontierCode 2.0 - ScoreCode47.320.2−27.1
BrokenArXiv 09/2026 (Native Harnesses)Math76.749.7−27.0
DAYJOB: AccountingFinance & Accounting34.47.6−26.8
ALE-BenchCode27482012−26.1
BuseyBench SVGMultimodal6.84.3−25.0
AA-PersonalityBench Conviction indexChat & Writing69.594.3+25.0
GameBench 2Games94.869.9−24.9
FrontierPhysics v0.1Science27.43.6−23.8
ZeroBench Main Mean pass@1 (five samples)Multimodal35.011.8−23.2
LAB-Bench LitQA2 (SecureBio)Science71.149.0−22.1
Mystery Game PuzzlesGames71.049.0−22.0
VCT v2 (SecureBio)Science60.538.5−22.0
Chess Puzzles (task version 2.0)Games67.046.0−21.0
AA-Omniscience IndexIndexes46.44.7−20.9
GBENCH One-shot Coding GScoreGames73.953.8−20.1
FrontierMath Tiers 1-3 (v2) - PrivateMath91.271.2−20.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Funds Asset ManagementLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - ImmigrationLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Intellectual PropertyLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Litigation Dispute ResolutionLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.020.0+20.0
ReactBench (39 tasks)Code54.434.4−20.0
ZeroBench (Main questions, pass@5)Multimodal43.023.0−20.0
Android Bench 2.0Code37.317.3−20.0
WARC-Bench (192-task test split)Agents90.370.5−19.8
LMArena Code Arena WebDevCode18141664−19.7
LMArena Agent Arena Praise vs ComplaintAgents40.31.2−19.6
AA-PersonalityBench Slop indexChat & Writing20.039.1+19.2
AA-PersonalityBench Informativeness indexChat & Writing90.571.4−19.2
LisanBench Difficulty-WeightedGames59637733+19.0
DAYJOB: LawLegal25.16.2−18.9
VCT (SecureBio)Science61.442.9−18.5
APEX-SWECode67.649.1−18.5
DAXBench v2Code95.777.7−18.1
VulcanBench Frontier v4 (functional pass@1)Code73.991.3+17.4
EnterpriseBench: CoreCraftProfessional Work74.458.5−15.9
Bio Capabilities Index (SecureBio)Indexes171149−15.8
Design Arena Overall Code CategoriesCode13841289−15.8
ARC-AGI-2General QA93.378.1−15.3
MBCT (SecureBio)Science68.753.9−14.8
AA-Briefcase v1.1 EloAgents16821416−14.6
Humanity's Sixth SenseMultimodal44.630.1−14.5
Hebbia Financial Services BenchmarkFinance & Accounting74.660.3−14.3
Humanity's Last Exam (Text-Only)General QA57.543.3−14.2
Excel Modeling BenchmarkFinance & Accounting75.961.8−14.1
Harvey HLAB — Criteria Pass RateLegal75.889.7+14.0
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.041.6−13.4
LAB-Bench ProtocolQA (SecureBio)Science94.081.9−12.1
GDP.pdf (Artificial Analysis, All-pass)Professional Work28.817.2−11.6
SciCode 1.0.1 (Artificial Analysis)Code66.955.9−11.0
CursorBench 4.0Code57.847.7−10.1
AA-PersonalityBench Organization indexChat & Writing79.188.6+9.6
CWE-bench v2 Pass@1Cybersecurity67.157.6−9.5
τ³-Banking (Fireworks, banking_knowledge)Agents56.046.8−9.2
LMArena Agent Arena Confirmed SuccessAgents20.72.3−9.2
LisanBench Path LengthGames2402921107−9.1
FrontierCode 1.1 Main (Artificial Analysis)Code52.243.7−8.5
WMDP-Bio (SecureBio)Science96.187.9−8.2
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.818.8+8.0
DAYJOB: TaxFinance & Accounting9.21.2−8.0
PubMedQA (SecureBio)Health & Medicine87.980.0−7.9
CritPt (avg@5)Science31.724.0−7.7
Epoch Capabilities Index (ECI)Indexes169159−7.5
Agents' Last Exam v1.1 (ALE v1.1)Agents43.436.1−7.4
LiveBench 2.0Agents86.579.2−7.3
LMArena Agent Arena Steering BurdenAgents18.64.0−7.3
MMMU ProMultimodal86.679.7−6.9
PI OfficeQA Pro v2Professional Work85.678.8−6.8
Excel Modeling Benchmark v1.1Finance & Accounting75.869.2−6.7
GDPxProfessional Work12.05.4−6.6
LMArena Agent Arena Bash ReliabilityAgents15.12.4−6.4
o11y-bench 1.1 Pass^3Agents87.581.3−6.2
APEX-Accounting 1.1Finance & Accounting65.559.3−6.2
Chartography (with tools)Multimodal89.083.0−6.0
Legacy-Bench (Fireworks, avg@3)Code62.056.3−5.7
Ruby LLM Benchmarks — Program FixerCode77.572.0−5.6
HLAB-Vals-1.1Legal9.214.2+5.0
WCB (SecureBio)Science54.549.6−4.9
Visa Agentic Commerce BenchmarkAgents62.257.8−4.4
PI-AutomationBench direct modeAgents92.896.8+4.0
ARC-AGI-1 Semi-PrivateGeneral QA98.594.5−4.0
AA-PersonalityBench Transparency indexChat & Writing90.586.7−3.8
DeepsecBench — Revalidated V1 (233 issues)Code26.630.0+3.4
VibeVideoBench (overall quality, no video = 0)Code4.13.9−3.3
GPQA-Bio (SecureBio)Science90.987.8−3.1
RedlineBenchLegal42.044.9+2.9
DeepSWE v1.1Code74.277.0+2.8
APEX-1 (Mercor, mean score)Professional Work65.868.3+2.5
Bug Hunt BenchCode39.737.3−2.4
STEMBench120 v1.0.0 (closed-book)Science2.62.5−2.1
AA-LCR v1.1Search84.782.7−2.0
MMLU-Bio (SecureBio)Science92.790.7−2.0
APEX-Agents 1.1Professional Work81.379.4−1.9
Finance Agent v2Finance & Accounting58.657.0−1.6
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine27.826.7−1.1
Access to Justice BenchmarkLegal73.873.1−0.7
OTIS Mock AIME 2024-2025Math100.099.4−0.6
LMArena Agent Arena Tool HallucinationAgents0.70.3−0.2
Epiq LAER CorporateBenchProfessional Work75.475.2−0.2
DualEntry Accounting AI BenchmarkFinance & Accounting82.282.20.0
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 - Data Privacy CybersecurityLegal0.00.00.0
Harvey's Legal Agent Benchmark - Employment LaborLegal0.00.00.0
Harvey's Legal Agent Benchmark - Energy Natural ResourcesLegal20.020.00.0
Harvey's Legal Agent Benchmark - Environmental ESGLegal0.00.00.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.020.00.0
Harvey's Legal Agent Benchmark - InsuranceLegal0.00.00.0
Harvey's Legal Agent Benchmark - International Trade SanctionsLegal0.00.00.0
Harvey's Legal Agent Benchmark - Real EstateLegal0.00.00.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.00.00.0
Harvey's Legal Agent Benchmark - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0

Claude Opus 5.5vsDeepSeek V4.1 Flash 1017normalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, SecureBio and 48 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…