Gemini 3.8 Flash vs Hy4
Hy4 leads by 1.2 capability points (≈95% interval −1.3 to +3.7): statistically tied · P(higher) ≈ 83%
Gemini 3.8 Flash wins 10 of 19 shared benchmarks (8 lost, 1 tied)
Capability vs. relative cost
Gemini 3.8 FlashHy4Closed weightsOpen weights
Hy4 has no setting with a relative cost yet.
Benchmark wins10 – 8 · 19 shared, 1 tied
Comparing matching setups only (144 benchmarks excluded) · suites counted once
Area← Gemini 3.8 FlashHy4 →
Data: Surge AI, Mercor, Neel Nanda and 7 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Hy4 minus Gemini 3.8 Flash.
| Chartography (November 2026 rescoring) | Perception & media › Documents & charts | 81.7 | 49.2 | −32.5 |
| Terminal-Bench 4.0 | Software › Software engineering | 19.1 | 44.4 | +25.3 |
| LLM Debate Benchmark | Language & writing › Language & wordplay | 1488 | 1647 | +22.7 |
| NoCoT-Bench NCRI 15.2 | Knowledge & reasoning › Abstract reasoning | 126 | 94.0 | −17.4 |
| APEX-Agents 1.1 | Professional › General knowledge work | 72.9 | 85.1 | +12.2 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 58.5 | 68.2 | +9.7 |
| Design Arena Overall Code Categories | Software › Frontend & design | 1299 | 1260 | −6.5 |
| NoCoT-Bench NCKI (kspine v3) | Knowledge & reasoning › General knowledge | 121 | 111 | −6.4 |
| RiemannBench (November 2026 rescoring) | Math | 56.8 | 51.2 | −5.6 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 2.0 | 7.2 | +5.2 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 4.5 | 9.2 | +4.7 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 63.2 | 66.8 | +3.6 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 26.5 | 23.7 | −2.8 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 11.4 | 13.6 | +2.2 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 34.3 | 32.7 | −1.6 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 48.8 | 48.1 | −0.7 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 44.2 | 43.6 | −0.6 |
| GDP.xlsx | Professional › General knowledge work | 18.3 | 17.7 | −0.6 |
| DAXBench v2 | Software › Data & ML | 87.2 | 87.2 | 0.0 |
Data: Surge AI, Mercor, Neel Nanda and 7 othersatlas.village.dev