Atlas

Compare Models

Claude Haiku 5.5 vs Muse Spark 1.3

Claude Haiku 5.5 leads by 0.1 capability points (≈95% interval −1.9 to +2.2): statistically tied · P(higher) ≈ 55%

Muse Spark 1.3 wins 79 of 127 shared benchmarks (44 lost, 4 tied)

Capability vs. relative cost
Claude Haiku 5.5Muse Spark 1.3Closed 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 5.5vsMuse Spark 1.3capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins44 – 79 · 127 shared, 4 tied

Comparing matching setups only (26 benchmarks excluded) · suites counted once
← Claude Haiku 5.5Muse Spark 1.3 →

Claude Haiku 5.5vsMuse Spark 1.3wins by categoryData Nov 15, 2026Data: Artificial Analysis, Vals AI, Mercor and 32 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Muse Spark 1.3 minus Claude Haiku 5.5.

AA-PersonalityBench Slop indexChat & Writing10.575.3+65.4
AA-PersonalityBench Transparency indexChat & Writing77.241.0−36.5
AA-PersonalityBench Imagination indexChat & Writing63.898.1+34.6
ProofBench v1.1Math89.058.0−31.0
AA-PersonalityBench Warmth indexChat & Writing69.596.2+26.9
AA-PersonalityBench Informativeness indexChat & Writing75.3100+25.0
AA-PersonalityBench Organization indexChat & Writing41.063.8+23.1
Harvey's Legal Agent BenchmarkLegal1.323.8+22.5
AutomationBench-AA 1.0.6Agents36.057.9+21.9
Medical Long Context Reasoning (MLCR-AA)Health & Medicine63.343.3−20.0
Extended NYT Connections 940 (Quadratic v1)Games65.785.1+19.4
ChessBench EloGames8751201+19.4
FrontierML ($200 wallet)Agents34.916.0−19.0
NoCoT-Bench NCKI (kspine v3)General QA85.9113+18.6
Chess Puzzles (task version 2.0)Games55.037.0−18.0
FrontierSWE V2Code43.825.8−18.0
Roboflow Vision Evals — Detection mAP@75 (September)Multimodal54.336.9−17.4
Roboflow Vision Evals — Detection mAP@50:95 (September)Multimodal53.035.8−17.2
Eval Connections One-Shot with TrapsGames67.084.0+17.0
NoCoT-Bench NCRI 15.2General QA68.398.3+16.4
AA-AnalystAgent (pass^5)Professional Work53.837.5−16.2
APEX-SWECode55.139.1−16.0
Code2Video-AA EloCode11921019−14.1
DistressBenchSafety71.784.9+13.2
LMArena Text Arena EloChat & Writing14221496+12.9
Artificial Analysis Coding Agent Index v1.5Indexes41.454.3+12.9
Roboflow Vision Evals — Identification, Strict (September)Multimodal80.292.7+12.5
FrontierCode 2.0 (Artificial Analysis)Code34.923.1−11.8
SpeciEval (Pooled English Runs)Safety86.998.7+11.8
MedXpertQA MM Extended (Mercor held-out)Health & Medicine48.460.1+11.7
Terminal-Bench 4.0Code36.424.7−11.6
Legal Research BenchLegal43.855.3+11.5
BuseyBench SVGMultimodal4.65.7+11.5
KernelBench CUDA Normalized (RTX PRO 6000)Code41.430.3−11.1
ChessBench CoherenceGames80.791.2+10.5
Tax Agent BenchFinance & Accounting62.272.4+10.2
Bug Hunt BenchCode20.530.7+10.2
GDPval Extended (Mercor held-out rubric score)Professional Work66.055.9−10.1
Harvey LAB (Mercor run, undisclosed ~213-task cohort)Legal4.915.0+10.1
MineBench — Bradley-Terry RatingGames15821747+10.0
ChessBench AccuracyGames66.976.0+9.8
AA-PersonalityBench Conviction indexChat & Writing67.677.2+9.6
Roboflow Vision Evals — Detection mAP@50 (September)Multimodal68.258.6−9.6
IOI (Vals v2)Code47.356.6+9.3
τ³-Banking (Mercor run, bm25_grep retrieval)Agents43.052.2+9.2
Terminal-Bench-Science 0.1Science20.011.0−9.0
AutomationBench 1.0.6 (Zapier)Agents12.020.7+8.7
MedXpertQA MM (200-question set)Health & Medicine67.876.3+8.5
Excel Modeling BenchmarkFinance & Accounting59.067.4+8.4
Harvey LAB-AA v1.1 (hallucination-gated)Legal0.68.9+8.3
APEX-1 (Mercor, mean score)Professional Work64.773.0+8.3
LMArena Code Arena WebDevCode15941655+8.0
DeepSWE v1.1Code59.967.8+7.9
AA-Omniscience IndexIndexes10.725.0+7.2
APEX-Agents 1.1Professional Work68.361.3−7.0
FrontierCode 1.1 Main (Artificial Analysis)Code45.538.6−6.8
CursorBench 4.0Code48.441.6−6.8
GameBench 2Games75.168.5−6.6
ACCoRDGeneral QA81.575.1−6.4
Roboflow Vision Evals — Identification, Judged (September)Multimodal86.592.7+6.2
KernelBench Mega Normalized (RTX PRO 6000)Code0.06.2+6.2
Humanity's Last Exam (Text-Only)General QA42.748.7+6.0
Finance Agent v2Finance & Accounting54.160.0+5.8
GDP.pdf (Artificial Analysis, All-pass)Professional Work20.826.6+5.8
AA-PersonalityBench Expressiveness indexChat & Writing77.271.4−5.8
VoxelBench Text RatingGames16961616−5.7
Decision-Bench v1.1Professional Work25.319.8−5.5
Mystery Game PuzzlesGames30.025.0−5.0
AA-3D v1.0 EloCode9981032+4.8
SciCode 1.0.1 (Artificial Analysis)Code55.059.7+4.7
Artificial Analysis Intelligence Index v4.3.2Indexes43.448.1+4.7
AA-LCR v1.1Search78.383.0+4.7
Vibe Code Bench v1.1Code90.485.9−4.6
Humanity's Last Exam DiamondGeneral QA29.633.7+4.1
CritPt v2 (agentic, avg@5)Science64.068.0+4.0
TaxCalcBench TY25Finance & Accounting2.06.0+4.0
Terminal-Bench 2.1Code85.088.8+3.8
SWE-Marathon v1.1Code50.046.3−3.7
Vals Index v2.1Indexes54.458.2+3.7
CritPt (avg@5)Science22.626.0+3.4
LMArena Agent Arena Net Improvement (Steering Burden)Agents0.52.2+3.2
Immigration BenchLegal61.865.0+3.2
DAXBench v2Code81.985.1+3.2
AA-LCR Extended (Mercor held-out)Search63.866.9+3.1
DeepFinanceFinance & Accounting40.037.0−3.0
GBENCH One-shot Coding GScoreGames55.652.7−2.9
Code MigrationCode44.647.4+2.8
LMCAGeneral QA51.153.9+2.8
CyberBench v1.1 OverallCybersecurity75.572.7−2.7
APEX-Accounting 1.1Finance & Accounting62.264.9+2.7
CVDP (517-problem set)Code57.054.3−2.7
Roboflow Vision Evals — Counting, Judged (September)Multimodal73.075.7+2.7
LMArena Agent Arena Bash ReliabilityAgents7.42.0−2.7
LMArena Agent Arena Praise vs ComplaintAgents-3.51.7+2.6
GDPval-AA v2.1Professional Work16291681+2.3
SRE Bench — Fully SolvedCode5.73.8−1.9
MMMU-Pro Extended (Mercor held-out)Multimodal31.933.8+1.9
Epoch Capabilities Index (ECI)Indexes160158−1.8
Roboflow Vision Evals — Counting, Strict (September)Multimodal73.074.8+1.8
DTBenchGeneral QA95.096.5+1.5
Roboflow Vision Evals — Reasoning, Judged (September)Multimodal74.873.3−1.5
LMArena Agent Arena Tool HallucinationAgents-2.10.8+1.4
Conceptual Reasoning IndexIndexes65.366.2+1.3
CharXiv Extended (Mercor held-out)Multimodal82.981.8−1.1
ForecastBench Preliminary Dataset Brier IndexForecasting60.861.9+1.1
LMArena Agent Arena Confirmed SuccessAgents1.63.7+1.1
MazeBench ASCIIGames1.00.0−1.0
MysteryMechanismScience35.136.0+0.9
LiveBench 2.0Agents75.674.8−0.9
SWE-bench Multimodal (Vals v1, 480 Tasks)Code44.043.1−0.8
FrontierMath Tiers 1-3 (v2) - PrivateMath75.174.4−0.7
SWE-bench Verified Extended (Mercor held-out)Code53.354.0+0.7
Artificial Analysis Progress Index v1Indexes326324−0.6
ALE-BenchCode19961981−0.5
Terminal-Bench 4.0 (Artificial Analysis)Code32.833.3+0.5
Roboflow Vision Evals — Reasoning, Strict (September)Multimodal72.873.3+0.5
AA-Briefcase v1.1 EloAgents15731581+0.5
OTIS Mock AIME 2024-2025Math98.999.2+0.3
CharXiv (descriptive and reasoning combined)Multimodal94.193.9−0.2
Artificial Analysis Intelligence Index v5.0Indexes36.236.4+0.2
Ramp Accounting BenchFinance & Accounting42.342.5+0.2
LMArena Agent Arena Steering BurdenAgents2.42.6+0.1
Artificial Analysis Intelligence Index v5.0.1Indexes36.536.6+0.1
AIME 2025Math100.0100.00.0
FrontierMath Tier 4 (v2) - PrivateMath46.346.30.0
ProgramBenchCode2.52.50.0
Terminal-Bench Science 0.1 (Vals)Science10.010.00.0

Claude Haiku 5.5vsMuse Spark 1.3normalised score marginsData Nov 15, 2026Data: Artificial Analysis, Vals AI, Mercor and 32 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…