Atlas

Compare Models

GPT-5.6 Terra vs Qwen3.8-Max

GPT-5.6 Terra leads by 1.0 capability points (≈95% interval −1.3 to +3.3): statistically tied · P(higher) ≈ 80%

Wins 24 of 34 shared benchmarks (10 lost)

Capability vs. relative cost
GPT-5.6 TerraQwen3.8-MaxClosed 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.

GPT-5.6 TerravsQwen3.8-Maxcapability vs. relative costData Nov 15, 2026Village LLM Atlasatlas.village.dev

Benchmark wins24 – 10 · 34 shared

Comparing matching setups only (306 benchmarks excluded) · suites counted once
← GPT-5.6 TerraQwen3.8-Max →

GPT-5.6 TerravsQwen3.8-Maxwins by categoryData Nov 15, 2026Data: Artificial Analysis, Conceptual Reasoning Benchmark Authors, Vercel and 9 othersVillage LLM Atlasatlas.village.dev

Margins

Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Qwen3.8-Max minus GPT-5.6 Terra.

YC-BenchGames6361442095838+51.0
BabaIsBench Stage 0: The IntroGames88.058.0−30.0
Terminal-Bench 4.0 (Artificial Analysis)Code35.418.7−16.7
Medical Long Context Reasoning (MLCR-AA)Health & Medicine31.719.4−12.2
MRCR v2 (8-needle, Context Arena) - 128K pointwiseGeneral QA80.292.3+12.1
𝜏³-BankingAgents40.251.3+11.1
AutomationBench-AA 1.0.6Agents59.649.2−10.4
CritPt (avg@5)Science30.020.0−10.0
LMCAGeneral QA55.046.2−8.7
Conceptual Reasoning IndexIndexes65.759.4−8.7
GDPval-AA v2Professional Work14791630+7.5
Terminal-Bench 2.1Code88.081.3−6.7
MRCR v2 (8-needle, Context Arena) - 128K AUCGeneral QA89.496.1+6.7
GDPval-AA v2.1Professional Work14611611+6.6
BuseyBench SVGMultimodal6.15.4−6.6
GDP.pdf (Artificial Analysis, All-pass)Professional Work24.620.2−4.4
DeepsecBench — Revalidated V1Code19.115.5−3.6
DeepsecBench — Revalidated V1 (233 issues)Code19.115.5−3.6
ACCoRDGeneral QA74.771.4−3.2
AA-Briefcase EloAgents13371388+2.9
Epoch Capabilities Index (ECI)Indexes161157−2.8
DeepsecBenchCode19.216.5−2.7
Continual Learning Exam (Trial 1)Agents27.825.7−2.2
GameBench 2Games53.551.4−2.0
Artificial Analysis Intelligence Index v4.3Indexes42.340.3−1.9
Artificial Analysis Intelligence Index v4.3.2Indexes42.140.2−1.9
SciCode 1.0.1 (Artificial Analysis)Code55.053.2−1.7
AA-Omniscience IndexIndexes0.13.4+1.7
MMMU ProMultimodal80.782.3+1.6
Artificial Analysis Progress Index v1Indexes324320−1.5
DTBenchGeneral QA93.392.0−1.3
Humanity's Last Exam (Text-Only)General QA41.943.0+1.2
AA-LCR v1.1Search79.078.3−0.7
GPQA DiamondGeneral QA92.992.6−0.3

GPT-5.6 TerravsQwen3.8-Maxnormalised score marginsData Nov 15, 2026Data: Artificial Analysis, Conceptual Reasoning Benchmark Authors, Vercel and 9 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…