Atlas

Compare Models

Claude Opus 5.5 vs DeepSeek V4.1 Flash

Claude Opus 5.5 leads by 8.3 capability points (≈95% interval +6.0 to +10.7): a clear lead over DeepSeek V4.1 Flash · P(higher) ≈ 100%

Wins 174 of 230 shared benchmarks (34 lost, 22 tied)

Capability vs. relative cost
Claude Opus 5.5DeepSeek V4.1 FlashClosed 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 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins174 – 34 · 230 shared, 22 tied

Comparing matching setups only (20 benchmarks excluded) · suites counted once
← Claude Opus 5.5DeepSeek V4.1 Flash →

Claude Opus 5.5vsDeepSeek V4.1 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 56 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 minus Claude Opus 5.5.

FrontierMath Tier 4 (v2) - PrivateMath95.026.8−68.2
BrokenArXiv 05/2026Math95.028.5−66.5
AA-PersonalityBench Imagination indexChat & Writing94.331.5−63.5
BrokenArXiv 06/2026Math100.039.4−60.6
Concept Synth INDUCTION Challenge64 Round 1General QA71.917.2−54.7
Concept Synth INDUCTION Challenge100 Round 1General QA65.011.0−54.0
LMArena Agent Arena Chat Net Improvement (Steering Burden)Agents23.31.6−53.0
Terminal-Bench-Science 0.1Science61.99.0−52.9
KernelBench Mega Normalized (RTX PRO 6000)Code100.048.2−51.8
Furniture Assembly BenchmarkMultimodal83.334.2−49.2
LLM Creative Story WritingChat & Writing3.7-0.5−47.5
ProofBench v1.1Math100.054.0−46.0
LMArena Agent Arena Code Net Improvement (Steering Burden)Agents22.92.7−45.6
RiemannBenchMath69.624.0−45.6
Terminal-Bench Science 0.1 (Vals)Science48.64.3−44.3
AA-PersonalityBench Warmth indexChat & Writing84.841.0−44.2
Medical Long Context Reasoning (MLCR-AA)Health & Medicine66.722.8−43.9
YC-BenchGames28737251657236−42.5
ArXivMath 08/2026 (Native Harnesses)Math86.043.9−42.1
LMArena Agent Arena Work Net Improvement (Steering Burden)Agents19.82.5−41.9
ArXivLean 09/2026Math47.25.7−41.5
ArXivMath 09/2026 (Native Harnesses)Math82.140.6−41.5
Code2Video-AA EloCode15361053−39.2
TaxCalcBench TY25Finance & Accounting42.04.0−38.0
Humanity's Last Exam DiamondGeneral QA54.619.1−35.5
Decision-Bench v1.1Professional Work59.125.3−33.9
Terminal-Bench 4.0Code64.831.2−33.6
TxBench Oligo Discovery V1 (113 evaluations)Science1.534.8+33.3
TCSExam LiteMath72.339.0−33.3
EnterpriseBench: CoreCraftProfessional Work74.441.5−32.9
Terminal-Bench 4.0 (Artificial Analysis)Code59.626.8−32.8
BioSecBench-Function V1 Clean RerunsScience73.640.8−32.8
ALE-BenchCode27481829−32.6
TxBench Antibody Discovery V1 (100 evaluations)Science8.339.7+31.3
APEX-Agents 1.1Professional Work81.350.5−30.8
AA-PersonalityBench Transparency indexChat & Writing90.560.0−30.8
FrontierCUA pass@1Agents36.06.7−29.3
SimpleQA Verified (Epoch anti-abstention)General QA72.242.9−29.3
KernelBench CUDA Normalized (RTX PRO 6000)Code76.347.0−29.3
OpenWorldEngineer (Mean@8)Code41.612.4−29.2
AA-PersonalityBench Expressiveness indexChat & Writing90.561.9−28.8
Mystery Game PuzzlesGames71.043.0−28.0
BuseyBench SVGMultimodal6.84.1−27.6
VIALSMultimodal40.212.6−27.6
SPU (Science Prediction Universe)Science-0.30.2+27.5
AA-PersonalityBench Conviction indexChat & Writing69.596.2+26.9
Conceptual Reasoning IndexIndexes77.958.4−26.7
GameBench 2Games94.868.4−26.3
MMMU-Pro Extended (Mercor held-out)Multimodal51.025.0−26.0
CloudOpsBench v1.0Agents47.821.8−26.0
GBENCH One-shot Coding GScoreGames73.948.0−25.9
AA-Omniscience IndexIndexes46.4-5.3−25.9
Creative Writing v3 EloChat & Writing20511540−25.8
SAFE-Teen 106 worst branch: Relationships (no prompt)Safety25.351.1−25.8
ArXivMath 06/2026Math83.357.6−25.7
LMArena Code Arena WebDevCode18141621−25.4
WeirdML v3Science31.25.9−25.3
LMArena Agent Arena Chat Praise vs ComplaintAgents51.93.1−24.4
CadQueryEval (Corrected Geometry)Code100.076.0−24.0
FrontierMath Tiers 1-3 (v2) - PrivateMath91.267.4−23.9
SWE-bench Multimodal (Vals v1, 480 Tasks)Code70.246.5−23.7
Design Creativity BenchCode51.043.6−23.5
ArXivMath 05/2026Math81.758.3−23.3
LMArena Agent Arena Code Praise vs ComplaintAgents49.73.0−23.3
SUPERChemScience79.656.7−22.9
MedXpertQA MM Extended (Mercor held-out)Health & Medicine71.248.4−22.8
BrokenArXiv 08/2026 (Native Harnesses)Math77.454.8−22.6
MedXpertQA MM (200-question set)Health & Medicine85.763.5−22.2
BrokenArXiv 09/2026 (Native Harnesses)Math76.754.7−22.0
SimpleBenchGeneral QA88.466.7−21.7
LMCAGeneral QA68.247.0−21.3
AA-AnalystAgent (pass^5)Professional Work56.335.0−21.2
AA-Briefcase v1.1 EloAgents18011415−21.2
AA-PersonalityBench Organization indexChat & Writing79.1100+21.2
ARC-AGI-2General QA93.372.9−20.4
ACCoRDGeneral QA88.468.1−20.3
TxBench Preclinical Pharmacology V1 Clean RerunsScience26.346.7+20.3
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science25.45.2−20.3
τ³-Banking (Fireworks, banking_knowledge)Agents56.035.7−20.3
DAXBench v2Code95.775.5−20.2
LLM Chess EloGames15751087−20.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Data Privacy CybersecurityLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.020.0+20.0
LMArena Agent Arena Work Praise vs ComplaintAgents43.63.7−19.9
VariantBench V1 Clean RerunsScience50.631.1−19.5
Bug Hunt BenchCode39.720.6−19.0
Chess Puzzles (task version 2.0)Games67.048.0−19.0
GeneBench-Pro Public Release SubsetScience30.011.0−19.0
MineBench — Bradley-Terry RatingGames21341822−18.9
Excel Modeling BenchmarkFinance & Accounting75.957.2−18.7
HealthBench ProfessionalHealth & Medicine71.452.9−18.5
OpenDS-WorldSearch38.020.5−17.5
SAFE-Teen 106 worst branch: Relationships (under-18 prompt)Safety2.419.8−17.5
CritPt (avg@5)Science31.714.3−17.4
AA-PersonalityBench Slop indexChat & Writing20.037.2+17.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.037.8−17.2
LLM Debate BenchmarkChat & Writing17011582−16.9
SWE-bench Verified Extended (Mercor held-out)Code66.082.7+16.7
SAFE-Teen 106 worst branch: Urgent Safety (under-18 prompt)Safety10.527.2−16.6
Senior SWE-BenchCode39.322.7−16.6
SciCode (main problems)Science27.311.2−16.1
GDP.pdf (Artificial Analysis, All-pass)Professional Work28.812.8−16.0
CharXiv Extended (Mercor held-out)Multimodal84.768.9−15.8
APEX-SWECode67.651.9−15.7
SWE-bench VerifiedCode98.282.9−15.3
SciCode 1.0.1 (Artificial Analysis)Code66.951.9−15.0
SAFE-Teen 106: Relationships (no prompt)Safety12.227.1−14.9
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine29.314.7−14.7
RuneBenchGames6.65.7−14.6
APEX-AccountingFinance & Accounting61.848.1−13.7
EpiBench V1 Clean RerunsScience39.025.5−13.5
LMArena Agent Arena Chat Confirmed SuccessAgents31.14.9−13.1
Eval Connections One-Shot with TrapsGames86.073.0−13.0
CWE-bench v1 Programmatic Pass@1Cybersecurity67.055.0−12.0
MacroscopeBenchCode83.972.0−11.9
GDPval-AA v2.1Professional Work18661600−11.7
Harvey HLAB — Criteria Pass RateLegal75.887.5+11.7
AA-PersonalityBench Informativeness indexChat & Writing90.579.1−11.5
Hebbia Financial Services BenchmarkFinance & Accounting74.663.1−11.5
Design Arena Overall Code CategoriesCode13841316−11.3
Westworld Finance Diligence BenchFinance & Accounting59.448.3−11.1
Longform Creative WritingChat & Writing84.674.0−10.6
LMArena Agent Arena Code Steering BurdenAgents22.82.3−10.2
SpeciEval (Refusal-Aware Retries, Historical)Safety87.297.4+10.1
SpeciEval (Pooled English Runs)Safety87.297.4+10.1
Epoch Capabilities Index (ECI)Indexes169156−9.9
Ramp Accounting BenchFinance & Accounting50.540.8−9.7
Creative Writing v3 Rubric ScoreChat & Writing16.014.1−9.7
SpatialBench V1 Clean RerunsScience72.863.2−9.6
SAFE-Teen 106 worst branch: Health Advice (under-18 prompt)Safety27.737.0−9.3
Terminal-Bench 2.1 Extended (Mercor held-out)Code47.538.4−9.1
BrowseComp (Mercor 130-question subset)Search88.579.4−9.1
DTBenchGeneral QA98.989.9−9.1
FrontierCode 1.1 Extended - Pass RateCode70.862.0−8.8
ITBench-AAAgents38.246.9+8.7
FrontierCode 1.1 Extended - ScoreCode65.356.6−8.7
MedCodeHealth & Medicine49.841.2−8.6
LiveBench 2.0Agents86.578.0−8.5
LMArena Agent Arena Work Steering BurdenAgents18.41.5−8.4
Arena Alignment IndexSafety83.274.8−8.4
LMArena Agent Arena Chat Bash ReliabilityAgents15.4-1.0−8.2
Excel Modeling Benchmark v1.1Finance & Accounting75.867.8−8.1
LMArena Agent Arena Code Confirmed SuccessAgents22.56.6−8.0
ScBench-Long V1 Clean RerunsScience6.814.8+8.0
Quant-Arena (iterative total score)Finance & Accounting44.336.5−7.8
LMArena Agent Arena Work Bash ReliabilityAgents15.50.1−7.7
LMArena Agent Arena Code Bash ReliabilityAgents15.1-0.2−7.6
BigFinanceBench private holdout (final-answer accuracy)Finance & Accounting48.240.6−7.6
BioSecBench-Refusal V1 Clean RerunsSafety3.811.3+7.5
LMArena Agent Arena Chat Steering BurdenAgents14.0-0.5−7.2
Terminal-Bench 2.1Code83.590.6+7.1
LMArena Agent Arena Work Confirmed SuccessAgents19.05.7−6.7
SWE Atlas Codebase QnA Extended (Mercor held-out)Code43.336.7−6.6
Immigration BenchLegal72.165.6−6.5
CharXiv (descriptive and reasoning combined)Multimodal96.490.1−6.3
SAFE-Teen 106: Health Advice (under-18 prompt)Safety17.323.6−6.3
BullshitBench v2 LeaderboardSafety63.057.0−6.0
SpatialBench-Long V1 Clean RerunsScience20.226.2+6.0
MedScribeHealth & Medicine91.485.5−5.9
SAGE v1.1Math91.385.4−5.9
LMArena Text Arena EloChat & Writing15061475−5.4
MMLU ProGeneral QA93.488.2−5.2
Finance Agent v2Finance & Accounting58.653.5−5.1
SAFE-Teen 106 worst branch: Urgent Safety (no prompt)Safety26.531.4−4.9
SAFE-Teen 106: Health Advice (no prompt)Safety35.030.2+4.8
SAFE-Teen 106: Urgent Safety (under-18 prompt)Safety3.17.9−4.7
SWE Atlas - Codebase QnACode48.953.5+4.6
BioSecBench-Surveillance V1 Clean RerunsScience55.851.5−4.3
Legacy-Bench (Fireworks, avg@3)Code62.057.7−4.3
DistressBenchSafety75.679.8+4.2
ARC-AGI-1 Semi-PrivateGeneral QA98.594.5−4.0
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine57.060.9+3.9
Humanity's Last Exam (with tools)General QA67.763.9−3.8
SAFE-Teen 106: Relationships (under-18 prompt)Safety0.44.2−3.7
LMArena Agent Arena Bash RecoveryAgents15.07.9−3.6
AA-LCR Extended (Mercor held-out)Search74.270.8−3.4
Financial Coaching BenchFinance & Accounting76.272.8−3.3
STEMBench120 v1.0.0 (closed-book)Science2.62.5−3.2
SAFE-Teen 106 worst branch: Health Advice (no prompt)Safety44.341.1+3.2
SWE-bench MultilingualCode95.598.2+2.7
Harvey LAB: In-House Contracting (held-out)Legal3.00.3−2.7
Harvey LAB-AA v1.1 (hallucination-gated)Legal4.21.7−2.5
dfbench Detection RecallCybersecurity54.957.3+2.4
HealthBenchHealth & Medicine64.361.9−2.4
SAGEMath45.847.9+2.0
DualEntry Accounting AI BenchmarkFinance & Accounting82.284.2+2.0
BioSecBench-Structure V1 (58 evaluations)Science29.327.6−1.7
AIME 2025Math100.098.3−1.7
Nimble Web ContextSearch71.873.5+1.7
OTIS Mock AIME 2024-2025Math100.098.3−1.7
GDPval (Mercor 213-task rubric score)Professional Work88.386.7−1.6
SAFE-Teen 106: Urgent Safety (no prompt)Safety12.711.3+1.4
MetagenomicsBench V1Science39.338.0−1.3
MetagenomicsBench V0Science39.338.0−1.3
AdvancedIFChat & Writing82.884.0+1.2
RCM-Bench-1Health & Medicine90.189.0−1.2
Extended NYT Connections 940 (Quadratic v1)Games88.589.6+1.1
Access to Justice BenchmarkLegal73.872.9−0.9
ScBench V1 Clean RerunsScience55.454.7−0.7
Harvey LAB held-out set (Fireworks run)Legal6.35.7−0.6
APEX-1 (Mercor, mean score)Professional Work65.865.5−0.3
ComplexConstraintsChat & Writing49.949.7−0.2
DeepResearch Bench II (Mercor 130-task set)Search53.353.2−0.1
Epiq LAER CorporateBenchProfessional Work75.475.3−0.1
LMArena Agent Arena Code Tool HallucinationAgents0.50.50.0
LMArena Agent Arena Chat Tool HallucinationAgents0.40.30.0
LMArena Agent Arena Work Tool HallucinationAgents0.50.50.0
DeepSWE v1.1Code74.274.20.0
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 - Emerging Companies Venture CapitalLegal0.00.00.0
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 - Funds Asset ManagementLegal0.00.00.0
Harvey's Legal Agent Benchmark - Healthcare Life SciencesLegal20.020.00.0
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 - 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 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Mercor, Artificial Analysis and 56 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…