Atlas

Compare Models

Claude Opus 5.5 vs Gemini 3.8 Flash

Claude Opus 5.5 leads by 7.1 capability points (≈95% interval +4.6 to +9.5): a clear lead over Gemini 3.8 Flash · P(higher) ≈ 100%

Wins 267 of 329 shared benchmarks (41 lost, 21 tied)

Capability vs. relative cost
Claude Opus 5.5Gemini 3.8 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 5.5vsGemini 3.8 Flashcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins267 – 41 · 329 shared, 21 tied

Comparing matching setups only (45 benchmarks excluded) · suites counted once
← Claude Opus 5.5Gemini 3.8 Flash →

Claude Opus 5.5vsGemini 3.8 Flashwins by categoryData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 89 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Gemini 3.8 Flash minus Claude Opus 5.5.

HPCT Refusal (CAIS)Safety100.013.0−87.0
VCT-Refusal (CAIS)Safety90.46.4−84.0
BrokenArXiv 06/2026Math100.020.8−79.2
BrokenArXiv 05/2026Math95.019.0−76.0
FrontierMath Tier 4 (v2) - PrivateMath95.022.0−73.0
Time Horizon Index: KSPAgents91.318.8−72.5
AA-PersonalityBench Transparency indexChat & Writing90.523.8−67.3
Odysseys2Agents71.210.7−60.5
Concept Synth INDUCTION Challenge64 Round 1General QA71.914.1−57.8
Concept Synth INDUCTION Challenge100 Round 1General QA65.09.0−56.0
AA-PersonalityBench Imagination indexChat & Writing94.341.0−53.8
FrontierML ($200 wallet)Agents58.85.7−53.1
ProofBench v1.1Math100.048.0−52.0
AA-PersonalityBench Slop indexChat & Writing20.071.4+51.9
Terminal-Bench 4.0Code64.813.1−51.7
Furniture Assembly BenchmarkMultimodal83.331.7−51.7
Code2Video-AA EloCode1536899−51.7
BrokenArXiv 09/2026 (Native Harnesses)Math76.726.7−50.0
BrokenArXiv 08/2026 (Native Harnesses)Math77.427.4−50.0
HoneyBench v0.2 (hack rate)Safety31.680.5−48.9
ViBench 2.0Code48.50.0−48.5
ArXivMath 08/2026 (Native Harnesses)Math86.040.4−45.6
Medical Long Context Reasoning (MLCR-AA)Health & Medicine66.721.7−45.0
Short Form Bench v1.0Games23891940−44.5
BioSecBench-Refusal V1 Clean RerunsSafety3.847.4+43.6
FrontierSWE V2 Worst@5Code52.09.5−42.5
LLM Creative Story WritingChat & Writing3.70.1−40.3
Harvey's Legal Agent Benchmark - Emerging Companies Venture CapitalLegal0.040.0+40.0
Terminal-Bench Science 0.1 (Vals)Science48.68.6−40.0
FrontierSWE V2 Best@5Code71.531.6−39.9
WitnessBench v2.0Games46.56.7−39.8
Hedge-BenchFinance & Accounting44.24.7−39.5
Terminal-Bench-Science 0.1Science49.010.0−39.0
MatFlowBench Pass@1Science56.718.9−37.8
BioSecBench-Function V1 Clean RerunsScience73.636.0−37.6
Humanity's Last Exam RMS Calibration Error (CAIS)General QA13.851.3−37.5
SAFE-Teen 106 worst branch: Relationships (no prompt)Safety25.362.6−37.4
Terminal-Bench 4.0 (Artificial Analysis)Code56.619.7−36.9
TxBench Antibody Discovery V1 Clean RerunsScience6.843.3+36.5
TaxCalcBench TY25Finance & Accounting42.06.0−36.0
CAD ArenaAgents74.641.0−33.5
AA-3D v1.0 EloCode1202974−33.1
LMArena Agent Arena Net Improvement (Steering Burden)Agents19.52.3−32.5
Decision-Bench v1.1Professional Work59.126.9−32.2
ArXivMath 09/2026 (Native Harnesses)Math82.150.0−32.1
TxBench Oligo Discovery V1 Clean RerunsScience1.433.1+31.7
APEX-SWECode67.636.4−31.2
MASKSafety89.458.6−30.8
GameBench 2Games94.864.0−30.8
Design Creativity BenchCode51.060.6+30.5
LMArena Code Arena WebDevCode18141583−30.3
sudo L7Code41.711.7−30.0
SAFE-Teen 106 worst branch: Urgent Safety (no prompt)Safety26.556.4−29.9
SWE-bench Multimodal (Vals v1, 480 Tasks)Code70.240.8−29.4
Olam Social Poker EloGames15061479−29.3
SAFE-Teen Relationships (no prompt)Safety13.041.7−28.6
DAYJOB: AccountingFinance & Accounting34.46.0−28.4
TxBench Preclinical Pharmacology V1 Clean RerunsScience26.354.7+28.3
CrossSearch (pass^3)Search57.029.0−28.0
SAFE-Teen 106: Relationships (no prompt)Safety12.240.1−27.9
FrontierPhysics v0.1Science27.40.0−27.4
Machiavelli Deception (CAIS)Safety77.586.5−27.3
LLM Debate BenchmarkChat & Writing17011514−26.6
SAFE-Teen Health Advice (no prompt)Safety32.859.1−26.3
AA-Briefcase v1.1 EloAgents16821204−26.2
FeatureArc (whole-feature chain gate)Code49.323.3−26.0
HANDBOOK.md Agents v1.1Professional Work52.326.5−25.8
VibeVideoBench (overall quality, no video = 0)Code4.12.8−25.5
AA-PersonalityBench Warmth indexChat & Writing84.860.0−25.0
SAFE-Teen 106 worst branch: Health Advice (no prompt)Safety44.369.0−24.7
SAFE-Teen 106: Health Advice (no prompt)Safety35.059.6−24.6
SAFE-Teen 106 worst branch: Urgent Safety (under-18 prompt)Safety10.534.8−24.3
Artificial Analysis Coding Agent Index v1.5Indexes66.041.9−24.1
Mystery Game PuzzlesGames71.047.0−24.0
DAYJOB: HealthcareHealth & Medicine24.70.8−23.9
WeirdML v3Science31.27.4−23.8
ArXivMath 05/2026Math81.758.3−23.3
MaxxBenchCode76.453.4−23.0
FrontierMath Tiers 1-3 (v2) - PrivateMath91.268.4−22.8
Bug Hunt BenchCode39.717.1−22.5
PLE (Physical-Engineering's Last Exam)Professional Work50.628.2−22.4
AURORA Multilingual τ³-bench (Turkish)Agents79.857.9−21.9
ScBench V1 Clean RerunsScience55.434.0−21.4
DAYJOB: FinanceFinance & Accounting23.92.8−21.1
SAFE-Teen Health Advice (under-18 prompt)Safety15.136.1−21.0
Biopharma Bench V0.1Health & Medicine67.546.6−20.9
Humanity's Last Exam DiamondGeneral QA55.034.3−20.7
EnigmaEvalGeneral QA57.837.2−20.6
Physical-Engineering's Last Exam (hidden test)Professional Work48.628.2−20.4
SAFE-Teen 106 worst branch: Health Advice (under-18 prompt)Safety27.748.1−20.4
SAFE-Teen 106: Health Advice (under-18 prompt)Safety17.337.5−20.2
SWE Atlas v1.1 - OverallCode77.457.3−20.1
ALE-BenchCode27482182−20.1
Harvey's Legal Agent Benchmark - Banking FinanceLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Capital MarketsLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Corporate GovernanceLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Structured Finance SecuritizationLegal0.020.0+20.0
Harvey's Legal Agent Benchmark - Trusts Estates Private ClientLegal0.020.0+20.0
MineBench — Bradley-Terry RatingGames21341806−19.9
MBCT (SecureBio)Science68.748.9−19.8
LAB-Bench LitQA2 (SecureBio)Science71.151.5−19.6
Vals RSI Index v1.1 (historical five tasks)Agents37.117.6−19.6
RSI-Exam 0.1 (Full 88)Agents0.50.3−19.5
VariantBench V1 Clean RerunsScience50.631.4−19.2
LiveBench 2.0Agents86.567.3−19.2
BrowseComp Extended (Mercor held-out)Search60.541.5−19.0
GeneBench-Pro Public Release SubsetScience30.011.0−19.0
ProgramBench (Mercor 100-task subset)Code20.01.0−19.0
SAFE-Teen 106: Urgent Safety (no prompt)Safety12.731.5−18.7
AA-LCR Extended (Mercor held-out)Search74.255.6−18.6
AURORA Multilingual τ³-bench (German)Agents79.060.5−18.5
RiemannBenchMath69.651.2−18.4
AURORA Multilingual τ³-bench (English)Agents81.162.8−18.3
BioSecBench-RespiratoryVirus-Surveillance V1 (110 evaluations)Science25.47.3−18.2
AURORA Multilingual τ³-bench (Korean)Agents78.460.5−17.9
SAFE-Teen Urgent Safety (no prompt)Safety13.131.0−17.9
LMArena Agent Arena Praise vs ComplaintAgents40.34.9−17.7
EpiBench V1 Clean RerunsScience39.021.4−17.6
SWE-bench VerifiedCode98.280.6−17.6
HiL-Bench (Human-in-Loop Benchmark)Agents59.041.5−17.5
Vals RSI Index v1.1 (four tasks)Agents37.319.9−17.4
HieroglyphBenchMultimodal37.454.6+17.2
PCT Political Inconsistency (CAIS)Safety29.146.3−17.2
FrontierCode 2.0 (Artificial Analysis)Code35.918.7−17.2
AURORA Multilingual τ³-bench (Persian)Agents77.760.6−17.1
MentalHealthBenchSafety52.435.5−16.9
Conceptual Reasoning IndexIndexes77.965.7−16.8
OncologyBench Strict (25 tasks, October 9 rubric)Health & Medicine29.313.3−16.0
Toloka Arena (v3)Agents75.059.0−16.0
ArXivMath 06/2026Math83.367.4−16.0
EnterpriseBench: CoreCraftProfessional Work74.458.5−15.9
Creative Writing v3 EloChat & Writing20511744−15.5
Remote Labor Index (RLI)Professional Work21.35.8−15.4
GDPval Extended (Mercor held-out rubric score)Professional Work75.860.4−15.4
Humanity's Last Exam (Text-Only)General QA57.542.1−15.4
AA-PersonalityBench Conviction indexChat & Writing69.584.8+15.4
LMCAGeneral QA68.252.9−15.3
VCT v2 (SecureBio)Science60.545.2−15.3
Clipping Bench v1.0Games2018596−15.3
GBENCH One-shot Coding GScoreGames73.958.6−15.2
BioTIER RefusalSafety98.783.9−14.8
MMMU-Pro Extended (Mercor held-out)Multimodal51.036.4−14.6
Hermes BenchAgents76.762.2−14.5
ChessBench CoherenceGames80.194.6+14.5
VCT (SecureBio)Science61.446.9−14.5
Harvey HLAB — Criteria Pass RateLegal75.890.2+14.4
ITBench-AAAgents38.252.5+14.3
Design Arena Overall Code CategoriesCode13841299−14.1
TreasuryQA BenchProfessional Work64.851.0−13.8
CritPt (avg@5)Science31.718.3−13.4
CVDP (517-problem set)Code64.150.7−13.4
SpatialBench V1 Clean RerunsScience72.859.4−13.3
SnorkelFinance 2.0Finance & Accounting18.65.4−13.2
SciCode (main problems)Science27.314.2−13.1
ScBench-Long V1 Clean RerunsScience6.819.9+13.1
MetagenomicsBench V1Science39.326.7−12.7
Blueprint-Bench 2Multimodal0.50.4−12.6
MetagenomicsBench V0Science39.326.7−12.6
AutoResearchExam (hidden-test AUARC)Agents0.60.5−12.6
Bio Capabilities Index (SecureBio)Indexes171153−12.5
HealthBench ProfessionalHealth & Medicine71.459.0−12.4
GDP.xlsxProfessional Work30.318.3−12.0
SWE-bench Verified Extended (Mercor held-out)Code66.054.0−12.0
SAFE-Teen Urgent Safety (under-18 prompt)Safety3.315.2−11.9
FrontierCode 1.1 Extended - ScoreCode65.353.5−11.8
SAFE-Teen 106: Urgent Safety (under-18 prompt)Safety3.114.9−11.8
GDPval-AA v2.1Professional Work17091442−11.8
BioSecBench-Surveillance V1 Clean RerunsScience55.844.1−11.7
SAFE-Teen 106 worst branch: Relationships (under-18 prompt)Safety2.414.0−11.6
SpatialViz-BenchMultimodal89.778.1−11.6
AA-PersonalityBench Expressiveness indexChat & Writing90.579.1−11.5
FrontierCode 1.1 Extended - Pass RateCode70.859.6−11.2
Veracode GenAI Code Security (Fall 2026)Cybersecurity73.062.0−11.0
CritPt v2 (agentic, avg@5)Science76.065.1−10.9
VIALSMultimodal40.229.4−10.8
SAGEMath45.835.1−10.8
ChessBench EloGames13491527+10.6
MedXpertQA MM Extended (Mercor held-out)Health & Medicine71.260.8−10.4
SciCode 1.0.1 (Artificial Analysis)Code66.956.6−10.3
Snorkel SWE-bench CLICode16.66.3−10.3
RuneBenchGames6.66.0−10.3
ACCoRDGeneral QA88.478.2−10.3
APEX-AccountingFinance & Accounting61.851.7−10.1
AA-AnalystAgent (pass^5)Professional Work56.346.3−10.0
LAB-Bench ProtocolQA (SecureBio)Science94.084.0−10.0
LMArena Agent Arena Steering BurdenAgents18.6-1.3−9.9
Financial Coaching BenchFinance & Accounting76.266.3−9.9
CUA-benchAgents14.04.2−9.8
Qodo AI Code Review Benchmark (125 PRs)Code56.847.0−9.8
Westworld Finance Diligence BenchFinance & Accounting59.449.7−9.7
SWE-bench MultilingualCode95.585.8−9.7
BenchBench-CloningScience20.510.8−9.7
LMArena Agent Arena Confirmed SuccessAgents20.71.3−9.7
AA-PersonalityBench Informativeness indexChat & Writing90.581.0−9.6
AML BenchFinance & Accounting66.457.1−9.3
MissionReportProfessional Work63.554.3−9.2
Chess Puzzles (task version 2.0)Games67.058.0−9.0
Extended NYT Connections 940 (Quadratic v1)Games88.597.4+8.9
Visa Agentic Commerce BenchmarkAgents62.253.3−8.9
Immigration BenchLegal72.163.4−8.7
FrontierCS Algorithmic v1.1Code56.547.8−8.7
HealthBench (length-adjusted, Artificial Analysis)Health & Medicine57.048.4−8.6
DAXBench v2Code95.787.2−8.5
AA-Omniscience IndexIndexes46.429.6−8.4
DAYJOB: TaxFinance & Accounting9.20.8−8.4
APEX-Agents 1.1Professional Work81.372.9−8.4
SpatialBench-Long V1 Clean RerunsScience20.228.6+8.3
SWE Atlas - Codebase QnACode48.940.6−8.3
Epoch Capabilities Index (ECI)Indexes169159−8.1
ChessBenchGames65.373.3+8.0
LMArena Agent Arena Bash RecoveryAgents15.0-0.7−7.8
Longform Creative WritingChat & Writing84.676.8−7.8
Snorkel Agentic Coding 2.0Code40.432.6−7.8
Arena Alignment IndexSafety83.275.8−7.4
HealthBenchHealth & Medicine64.357.1−7.1
SnorkelRevOpsProfessional Work12.55.4−7.1
SUPERChemScience79.672.6−7.0
MedScribeHealth & Medicine91.484.5−6.9
LegalReasonBenchLegal46.539.6−6.9
ATLAS-Finance v1.0Finance & Accounting11.44.7−6.7
Access to Justice BenchmarkLegal73.867.1−6.7
WMDP-Bio (SecureBio)Science96.189.4−6.7
Snorkel Digital Work IndexIndexes23.616.9−6.7
Cresta-BenchAgents82.676.1−6.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents55.048.8−6.2
SnorkelLegalLegal31.425.3−6.1
GDP.pdf (Artificial Analysis, All-pass)Professional Work28.822.8−6.0
SimpleBenchGeneral QA88.482.4−6.0
WellnessBenchSafety53.047.0−6.0
CharXiv Extended (Mercor held-out)Multimodal84.778.9−5.8
HardGeoBench Avg DistanceMultimodal655843−5.8
AA-PersonalityBench Organization indexChat & Writing79.173.3−5.8
IntPhys 2Multimodal68.362.5−5.7
BrowseComp (Mercor 130-question subset)Search88.583.1−5.4
Agent Red Teaming k=100 (CAIS)Safety24.429.8−5.4
micro1 PharmaBenchHealth & Medicine10.85.6−5.2
CloudOpsBench v1.0Agents47.842.8−5.0
DistressBenchSafety75.670.7−4.8
GDPval (Mercor 213-task rubric score)Professional Work88.383.5−4.8
Ramp Accounting BenchFinance & Accounting50.545.7−4.8
τ³-Banking (Fireworks, banking_knowledge)Agents56.051.4−4.6
SWE-TogetherCode68.864.2−4.6
ERQAMultimodal75.179.6+4.5
Gray Swan IPI Attack Success (k=15)Safety1.05.5−4.5
ARC-AGI-2General QA93.389.2−4.2
Terminal-Bench 2.1Code83.587.6+4.1
FACTS Benchmark SuiteGeneral QA65.369.4+4.1
CadQueryEval (Corrected Geometry)Code100.096.0−4.0
Excel Modeling BenchmarkFinance & Accounting75.972.2−3.7
SnorkelUnderwrite 2.0Finance & Accounting22.719.1−3.6
SAFE-Teen Relationships (under-18 prompt)Safety0.54.0−3.5
LVBenchMultimodal83.787.1+3.4
Harvey LAB: In-House Contracting (held-out)Legal3.06.3+3.3
SnorkelManufacturingProfessional Work10.47.1−3.3
AIME 2025Math100.096.7−3.3
SAFE-Teen 106: Relationships (under-18 prompt)Safety0.43.7−3.3
TextQuests Game ProgressGames57.354.1−3.2
DTBenchGeneral QA98.995.7−3.2
Snorkel WorkplaceAgentsProfessional Work14.411.2−3.2
Terminal-Bench 2.1 Extended (Mercor held-out)Code47.544.4−3.1
Ivo-micro1 ContractBenchLegal58.155.0−3.1
SAGE v1.1Math91.388.2−3.1
APEX-1 (Mercor, mean score)Professional Work65.868.8+3.0
Humanity's Sixth SenseMultimodal44.641.6−3.0
Excel Modeling Benchmark v1.1Finance & Accounting75.872.9−3.0
FrontierMath ErdősMath2.90.0−2.9
HLAB-Vals-1.1Legal9.212.1+2.9
MMLU ProGeneral QA93.490.7−2.7
SpeciEval (Refusal-Aware Retries, Historical)Safety87.289.9+2.6
SpeciEval (Pooled English Runs)Safety87.289.9+2.6
LMArena Agent Arena Bash ReliabilityAgents15.19.9−2.6
WCB (SecureBio)Science54.557.1+2.6
HardGeoBench <=100miMultimodal62.560.0−2.5
HardGeoBench <=10miMultimodal43.846.3+2.5
SimpleQA Verified (Epoch anti-abstention)General QA72.269.7−2.5
BuseyBench SVGMultimodal6.86.6−2.2
Roboflow Vision Evals (October 2026, five tasks)Multimodal82.079.8−2.2
Epiq LAER CorporateBenchProfessional Work75.477.6+2.2
MMMU ProMultimodal86.684.5−2.1
Eval Connections One-Shot with TrapsGames86.088.0+2.0
ChessBench AccuracyGames79.381.1+1.9
Creative Writing v3 Rubric ScoreChat & Writing16.015.6−1.9
ForecastBench Preliminary Dataset Brier IndexForecasting62.260.5−1.7
MedCodeHealth & Medicine49.848.1−1.7
ComplexConstraintsChat & Writing49.948.4−1.5
OpenDS-WorldSearch38.036.5−1.5
LMArena Text Arena EloChat & Writing15061497−1.5
SM BenchSafety81.880.4−1.3
Professional Reasoning Benchmark - LegalLegal47.648.9+1.3
HardGeoBench Country AccuracyMultimodal68.867.5−1.2
Quant-Arena (iterative total score)Finance & Accounting44.343.0−1.2
LLM Chess EloGames15751546−1.2
OTIS Mock AIME 2024-2025Math100.098.9−1.1
CharXiv (descriptive and reasoning combined)Multimodal96.495.4−1.0
DualEntry Accounting AI BenchmarkFinance & Accounting82.283.2+1.0
MedXpertQA MM (200-question set)Health & Medicine85.784.7−1.0
MindCube TinyMultimodal93.192.2−0.9
BioSimBench 1.0Science14.515.3+0.8
RedlineBenchLegal42.041.3−0.7
TextQuests Average Cumulative HarmSafety24.224.4−0.7
AA-LCR v1.1Search84.784.0−0.7
MacroscopeBenchCode70.170.7+0.6
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal10.811.4+0.6
Roboflow Vision Evals (September 2026 protocol)Multimodal85.585.1−0.4
DeepSWE v1.1Code74.273.8−0.4
Prophet Arena Arena Score (Fixed Context)Forecasting85.585.3−0.2
Prophet Arena Arena Score (Agentic, OpenCode)Forecasting86.286.0−0.2
LMArena Agent Arena Tool HallucinationAgents0.70.4−0.1
ARC-AGI-1 Semi-PrivateGeneral QA98.598.50.0
Dual-Arm Sim (task completion)Agents0.00.00.0
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 - Bankruptcy RestructuringLegal0.00.00.0
Harvey's Legal Agent Benchmark - Corporate M&ALegal0.00.00.0
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 - 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 - TaxLegal0.00.00.0
Harvey's Legal Agent Benchmark - White Collar Defense InvestigationsLegal0.00.00.0
ObviousBenchGeneral QA100.0100.00.0

Claude Opus 5.5vsGemini 3.8 Flashnormalised score marginsData Nov 15, 2026Data: Vals AI, Artificial Analysis, Mercor and 89 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…