Atlas

Compare Models

Grok 4.7 vs Qwen3.8 Max (0902)

Grok 4.7 leads by 0.3 capability points (≈95% interval −2.0 to +2.6): statistically tied · P(higher) ≈ 60%

Wins 34 of 52 shared benchmarks (17 lost, 1 tied)

Capability vs. relative cost
Grok 4.7Qwen3.8 Max (0902)Closed 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.

Grok 4.7vsQwen3.8 Max (0902)capability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins34 – 17 · 52 shared, 1 tied

Comparing matching setups only (179 benchmarks excluded) · suites counted once
← Grok 4.7Qwen3.8 Max (0902) →

Grok 4.7vsQwen3.8 Max (0902)wins by categoryData Nov 15, 2026Data: Artificial Analysis, AIMING Lab (UNC-Chapel Hill), BuseyBench and 7 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Qwen3.8 Max (0902) minus Grok 4.7.

AA-PersonalityBench Slop indexChat & Writing8.673.3+65.4
AA-PersonalityBench Expressiveness indexChat & Writing75.323.8−51.9
AA-PersonalityBench Conviction indexChat & Writing82.935.3−48.1
AA-PersonalityBench Organization indexChat & Writing29.671.4+42.3
AA-PersonalityBench Imagination indexChat & Writing35.348.6+13.5
AA-PersonalityBench Transparency indexChat & Writing58.144.8−13.5
Terminal-Bench 4.0 (Artificial Analysis)Code25.838.9+13.1
BuseyBench SVGMultimodal4.65.8+12.2
FrontierSWE V2Code29.517.8−11.7
CritPt v2 (agentic, avg@5)Science62.351.4−10.9
Terminal-Bench 4.0Code37.627.0−10.6
AA-Omniscience IndexIndexes32.012.0−10.0
AA-AnalystAgent (pass^5)Professional Work55.045.0−10.0
DAXBench v2Code92.683.0−9.6
AutomationBench-AA 1.0.6Agents65.656.2−9.4
Code2Video-AA EloCode11271021−8.6
Artificial Analysis Strategy & Ops IndexIndexes54.847.3−7.6
VIALS-AA (pass@3)Multimodal65.872.7+6.8
Artificial Analysis Finance & Accounting IndexIndexes51.745.0−6.7
VIALS-AA (pass^3)Multimodal37.931.7−6.2
Artificial Analysis Legal IndexIndexes47.241.0−6.2
AA-PersonalityBench Informativeness indexChat & Writing33.439.1+5.8
Artificial Analysis Healthcare & Medical IndexIndexes46.640.8−5.8
SciCode 1.0.1 (Artificial Analysis)Code57.852.1−5.7
AA-3D v1.0 EloCode992954−5.6
FrontierCode 2.0 (Artificial Analysis)Code28.723.4−5.2
Medical Long Context Reasoning (MLCR-AA)Health & Medicine15.020.0+5.0
MMMU ProMultimodal78.282.8+4.6
AA-PersonalityBench Warmth indexChat & Writing44.848.6+3.8
Terminal-Bench 2.1Code85.088.8+3.8
FrontierCode 1.1 Main (Artificial Analysis)Code39.643.2+3.6
Artificial Analysis Economics IndexIndexes45.742.5−3.2
AA-Briefcase v1.1 EloAgents16571604−2.9
MKA (Artificial Analysis)Professional Work40.137.4−2.7
Artificial Analysis Engineering IndexIndexes41.839.3−2.5
GDPval-AA v2.1Professional Work17181674−1.9
Artificial Analysis Intelligence Index v5.0.1Indexes35.533.7−1.8
ITBench-AAAgents42.140.3−1.8
Artificial Analysis Intelligence Index v5.0Indexes35.333.6−1.8
VIALS-AA (average of 3)Multimodal52.450.9−1.4
Artificial Analysis Intelligence Index v4.3.2Indexes46.445.4−1.0
Veracode GenAI Code Security (Fall 2026)Cybersecurity69.068.0−1.0
RSI-Exam 0.1 (Full 88)Agents0.40.4+0.8
Terminal-Bench-Science 0.1Science11.411.9+0.5
GDP.pdf (Artificial Analysis, All-pass)Professional Work23.222.8−0.4
Humanity's Last Exam DiamondGeneral QA23.624.0+0.4
Artificial Analysis Progress Index v1Indexes323324+0.3
AA-LCR v1.1Search80.080.3+0.3
CritPt (avg@5)Science18.017.7−0.3
Epoch Capabilities Index (ECI)Indexes156156−0.2
Humanity's Last Exam (Text-Only)General QA43.143.10.0
Visa Agentic Commerce BenchmarkAgents60.060.00.0

Grok 4.7vsQwen3.8 Max (0902)normalised score marginsData Nov 15, 2026Data: Artificial Analysis, AIMING Lab (UNC-Chapel Hill), BuseyBench and 7 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…