Atlas

Compare Models

Claude Haiku 5.5 vs DeepSeek V4.1 Flash 1017

Claude Haiku 5.5 leads by 0.4 capability points (≈95% interval −1.9 to +2.7): statistically tied · P(higher) ≈ 63%

DeepSeek V4.1 Flash 1017 wins 56 of 105 shared benchmarks (48 lost, 1 tied)

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

Benchmark wins48 – 56 · 105 shared, 1 tied

Comparing matching setups only (7 benchmarks excluded) · suites counted once
← Claude Haiku 5.5DeepSeek V4.1 Flash 1017 →

Claude Haiku 5.5vsDeepSeek V4.1 Flash 1017wins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 26 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 Haiku 5.5.

HardGeoBench Avg DistanceMultimodal13193201−58.1
LisanBench Difficulty-WeightedGames27117733+53.8
AA-PersonalityBench Organization indexChat & Writing41.088.6+48.1
AA-PersonalityBench Warmth indexChat & Writing69.527.7−42.3
HardGeoBench <=100miMultimodal45.07.5−37.5
HardGeoBench Country AccuracyMultimodal51.315.0−36.2
AutomationBench-AA 1.0.6Agents36.068.5+32.5
AA-PersonalityBench Slop indexChat & Writing10.539.1+28.8
AA-PersonalityBench Conviction indexChat & Writing67.694.3+26.9
AA-PersonalityBench Imagination indexChat & Writing63.837.2−26.9
HardGeoBench <=10miMultimodal30.03.8−26.2
LisanBench Path LengthGames1289721107+25.5
AA-PersonalityBench Expressiveness indexChat & Writing77.252.4−25.0
Mystery Game PuzzlesGames30.049.0+19.0
SimpleQA Verified (Epoch anti-abstention)General QA23.842.5+18.7
ARC-AGI-2General QA61.378.1+16.8
Bug Hunt BenchCode20.537.3+16.8
Code Migration v1.1Code52.467.5+15.2
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.918.8+13.9
Tax Agent BenchFinance & Accounting62.274.9+12.7
Cresta-BenchAgents20.132.6+12.6
FrontierML ($200 wallet)Agents34.923.5−11.5
APEX-Agents 1.1Professional Work68.379.4+11.1
NoCoT-Bench NCKI (kspine v3)General QA85.9102+10.9
CritPt v2 (agentic, avg@5)Science62.652.3−10.3
Artificial Analysis Strategy & Ops IndexIndexes39.149.2+10.1
HLAB-Vals-1.1Legal4.214.2+10.0
AA-PersonalityBench Transparency indexChat & Writing77.286.7+9.6
LMArena Code Arena WebDevCode15941664+9.1
Chess Puzzles (task version 2.0)Games55.046.0−9.0
FrontierCode 2.0 - Pass RateCode31.923.0−8.9
Humanity's Last Exam DiamondGeneral QA29.620.8−8.8
FrontierCode 2.0 - ScoreCode28.920.2−8.7
VIALS-AA (pass^3)Multimodal49.140.4−8.7
Medical Long Context Reasoning (MLCR-AA, October 2026 methodology)Health & Medicine35.026.7−8.3
Harvey's Legal Agent BenchmarkLegal1.39.2+7.9
Artificial Analysis Intelligence Index v5.0Indexes36.228.5−7.7
NoCoT-Bench NCRI 15.2General QA68.381.6+7.3
VIALS-AA (average of 3)Multimodal60.553.2−7.2
Terminal-Bench-Science 0.1Science20.012.9−7.1
APEX-SWECode55.149.1−6.0
DualEntry Accounting AI BenchmarkFinance & Accounting76.282.2+6.0
ARC-AGI-1 Semi-PrivateGeneral QA89.094.5+5.5
GameBench 2Games75.169.9−5.2
VIALS-AA (pass@3)Multimodal71.466.5−5.0
Artificial Analysis Engineering IndexIndexes42.537.6−4.9
Terminal-Bench 4.0Code36.431.8−4.5
VulcanBench Frontier v4 (functional pass@1)Code95.791.3−4.3
AA-LCR v1.1Search78.382.7+4.3
RedlineBenchLegal40.644.9+4.3
DAXBench v2Code81.977.7−4.3
Artificial Analysis Economics IndexIndexes42.738.5−4.2
Artificial Analysis Intelligence Index v5.0.1Indexes36.532.6−3.9
FrontierMath Tiers 1-3 (v2) - PrivateMath75.171.2−3.9
AA-PersonalityBench Informativeness indexChat & Writing75.371.4−3.8
Artificial Analysis Healthcare & Medical IndexIndexes37.541.3+3.8
Excel Modeling Benchmark v1.1Finance & Accounting65.569.2+3.7
GDP.pdf (Artificial Analysis, All-pass)Professional Work20.817.2−3.6
APEX-1 (Mercor, mean score)Professional Work64.768.3+3.6
LiveBench 2.0Agents75.679.2+3.6
Vals Finance IndexIndexes57.460.9+3.4
Wiz Cyber Model Arena v3.9Cybersecurity57.060.3+3.3
ARC-AGI-3 (Semi-Private)General QA2.85.8+3.0
AA-Omniscience IndexIndexes10.74.7−3.0
APEX-Accounting 1.1Finance & Accounting62.259.3−2.9
LMArena Agent Arena Net Improvement (Steering Burden)Agents0.52.0+2.9
Finance Agent v2Finance & Accounting54.157.0+2.8
Excel Modeling BenchmarkFinance & Accounting59.061.8+2.8
Vibe Code Bench v1.1Code90.487.7−2.7
Vals Index v2.2Indexes58.761.3+2.5
LMArena Agent Arena Bash ReliabilityAgents7.42.4−2.5
Finance Agent v2.1Finance & Accounting65.462.9−2.4
MKA (Artificial Analysis)Professional Work41.038.5−2.4
LMArena Agent Arena Praise vs ComplaintAgents-3.51.2+2.3
Vals Index v2.1Indexes54.456.6+2.1
BuseyBench SVGMultimodal4.64.3−2.1
GBENCH One-shot Coding GScoreGames55.653.8−1.9
FrontierCode 1.1 Main (Artificial Analysis)Code41.843.7+1.9
Artificial Analysis Intelligence Index v4.3.2Indexes43.445.1+1.7
Vals Legal IndexIndexes70.371.9+1.6
o11y-bench 1.1 Pass^3Agents82.881.3−1.6
Design Arena Overall Code CategoriesCode12801289+1.5
Legal Research BenchLegal43.845.2+1.4
CritPt (avg@5)Science22.624.0+1.4
Artificial Analysis Finance & Accounting IndexIndexes43.642.2−1.4
τ³-Banking (Mercor run, bm25_grep retrieval)Agents43.041.6−1.4
AA-Briefcase v1.1 EloAgents14411416−1.4
LMArena Agent Arena Tool HallucinationAgents-2.10.3+1.2
Terminal-Bench 4.0 (Artificial Analysis)Code32.833.8+1.0
Artificial Analysis Legal IndexIndexes38.737.7−0.9
SciCode 1.0.1 (Artificial Analysis)Code55.055.9+0.9
LMArena Agent Arena Steering BurdenAgents2.44.0+0.8
Artificial Analysis Progress Index v1Indexes326324−0.8
Epoch Capabilities Index (ECI)Indexes160159−0.7
CursorBench 4.0Code48.447.7−0.7
Humanity's Last Exam (Text-Only)General QA42.743.3+0.6
MMMU ProMultimodal79.179.7+0.6
FrontierPhysics v0.1Science4.23.6−0.6
ALE-BenchCode19962012+0.6
Political Science LLM BenchmarkProfessional Work65.766.2+0.6
OTIS Mock AIME 2024-2025Math98.999.4+0.6
GDPxProfessional Work4.95.4+0.5
LMArena Agent Arena Confirmed SuccessAgents1.62.3+0.4
Code MigrationCode44.644.2−0.4
FrontierMath Tier 4 (v2) - PrivateMath46.346.30.0

Claude Haiku 5.5vsDeepSeek V4.1 Flash 1017normalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Epoch AI and 26 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…