Gemini 3.8 Flash vs Gemini 4 Neon
Gemini 4 Neon leads by 8.3 capability points (≈95% interval +5.6 to +11.0): a clear lead over Gemini 3.8 Flash · P(higher) ≈ 100%
Wins 65 of 66 shared benchmarks (1 lost)
Capability vs. relative cost
Gemini 3.8 FlashGemini 4 NeonClosed weightsOpen weights
Benchmark wins1 – 65 · 66 shared
Comparing matching setups only (6 benchmarks excluded) · suites counted once
Area← Gemini 3.8 FlashGemini 4 Neon →
Data: Vals AI, Surge AI, Artificial Analysis and 10 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Gemini 4 Neon minus Gemini 3.8 Flash.
| FrontierMath Tier 4 (v2) - Private | Math | 22.0 | 97.6 | +75.6 |
| Vending-Bench 2 | Business simulations | 5094 | 16843 | +69.6 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 31.7 | 89.2 | +57.5 |
| Terminal-Bench-Science 0.1 | Science › Multidisciplinary | 10.0 | 65.7 | +55.7 |
| Terminal-Bench Science 0.1 (Vals) | Science › Multidisciplinary | 8.6 | 62.9 | +54.3 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 4.5 | 50.5 | +46.0 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 2.0 | 47.2 | +45.2 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 6.0 | 49.6 | +43.6 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 29.1 | 71.3 | +42.2 |
| Terminal-Bench 4.0 (Artificial Analysis) | Software › Software engineering | 19.7 | 61.6 | +41.9 |
| Terminal-Bench 4.0 | Software › Software engineering | 19.7 | 61.6 | +41.9 |
| Mystery Game Puzzles | Games | 47.0 | 85.0 | +38.0 |
| Code Migration v1.1 (Vals Index v2.2 subset) | Software › Software engineering | 38.2 | 75.8 | +37.6 |
| Bug Hunt Bench | Software › Software engineering | 17.1 | 54.3 | +37.1 |
| Humanity's Sixth Sense | Perception & media › Spatial & visual reasoning | 41.6 | 75.9 | +34.3 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 1.5 | 18.4 | +33.0 |
| Chartography | Perception & media › Documents & charts | 40.9 | 73.5 | +32.6 |
| Artificial Analysis Intelligence Index v5.0.1 | Composite indexes | 30.8 | 61.4 | +30.6 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 57.3 | 87.4 | +30.2 |
| sudo L7 (October 2026 rescoring) | Software › Software engineering | 10.6 | 38.3 | +27.7 |
| RiemannBench (November 2026 rescoring) | Math | 56.8 | 84.0 | +27.2 |
| Vals Science Index | Science › Multidisciplinary | 35.8 | 62.6 | +26.9 |
| AutomationBench 1.0.6 (Zapier) | Tool & computer use › Customer service & operations | 29.7 | 56.3 | +26.6 |
| DAYJOB: Law (November 2026 release) | Professional › Legal | 0.0 | 26.2 | +26.2 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 26.5 | 52.3 | +25.8 |
| Chess Puzzles (task version 2.0) | Games | 58.0 | 83.0 | +25.0 |
| GDP.xlsx | Professional › General knowledge work | 18.3 | 43.1 | +24.8 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 18.7 | 43.3 | +24.5 |
| LMArena Code Arena WebDev | Software › Frontend & design | 1585 | 1764 | +23.8 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1204 | 1632 | +23.5 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 68.4 | 91.2 | +22.8 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 4.0 | 44.5 | +20.2 |
| MKA (Artificial Analysis) | Professional › General knowledge work | 35.3 | 54.6 | +19.3 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 58.5 | 77.4 | +18.9 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 60.9 | 79.3 | +18.4 |
| BioMysteryBench (Vals Terminus 2) | Science › Life sciences | 62.2 | 80.0 | +17.8 |
| Legal Research Bench | Professional › Legal | 38.9 | 56.3 | +17.3 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 15.6 | 31.7 | +16.1 |
| APEX-Agents 1.1 | Professional › General knowledge work | 72.9 | 88.8 | +15.9 |
| AA-Omniscience v1.1 Index | Knowledge & reasoning › General knowledge | 23.7 | 54.8 | +15.5 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 11.4 | 26.6 | +15.2 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 78.7 | 93.7 | +15.0 |
| Vals Index v2.2 | Composite indexes | 56.0 | 70.9 | +14.9 |
| Vals RSI Index v1.1 (four tasks) | Software › Data & ML | 19.9 | 34.4 | +14.5 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 64.6 | 78.0 | +13.4 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 66.8 | 79.8 | +13.0 |
| APEX-Accounting 1.1 | Professional › Finance, accounting & tax | 58.8 | 70.8 | +12.0 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 0.8 | 12.8 | +12.0 |
| SimpleBench | Knowledge & reasoning › General knowledge | 82.4 | 94.4 | +12.0 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | -2.8 | 21.2 | +12.0 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 48.8 | 59.5 | +10.7 |
| GDP.pdf (Artificial Analysis, All-pass) | Professional › General knowledge work | 22.8 | 33.4 | +10.6 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 63.2 | 72.9 | +9.7 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 44.2 | 53.4 | +9.2 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 1.0 | 19.2 | +9.1 |
| Vals Legal Index | Professional › Legal | 69.8 | 78.4 | +8.6 |
| MysteryMechanism | Math | 36.5 | 45.0 | +8.6 |
| Vals Finance Index | Professional › Finance, accounting & tax | 60.5 | 68.3 | +7.9 |
| HLAB-Vals-1.1 | Professional › Legal | 12.1 | 19.6 | +7.5 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 8.9 | -6.0 | −7.4 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 69.7 | 75.2 | +5.5 |
| LMArena Text Arena Elo | Language & writing › General assistance | 1497 | 1518 | +3.8 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 72.9 | 76.5 | +3.6 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 65.8 | 68.6 | +2.8 |
| OTIS Mock AIME 2024-2025 | Math | 98.9 | 100.0 | +1.1 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 0.0 | 1.0 | +0.5 |
Data: Vals AI, Surge AI, Artificial Analysis and 10 othersatlas.village.dev