Claude Fable 5.2 vs Gemini 4 Neon
Claude Fable 5.2 leads by 0.8 capability points (≈95% interval −2.3 to +3.9): statistically tied · P(higher) ≈ 70%
Wins 29 of 56 shared benchmarks (26 lost, 1 tied)
Capability vs. relative cost
Claude Fable 5.2Gemini 4 NeonClosed weightsOpen weights
Benchmark wins29 – 26 · 56 shared, 1 tied
Comparing matching setups only (2 benchmarks excluded) · suites counted once
Area← Claude Fable 5.2Gemini 4 Neon →
Data: Surge AI, Vals AI, Epoch AI and 14 othersatlas.village.dev
Margins
Each benchmark scaled 0–100% (its declared range, else the field’s range); Δ is Gemini 4 Neon minus Claude Fable 5.2.
| Vending-Bench 2 | Business simulations | 9628 | 16843 | +42.8 |
| SWE-Marathon v1.2 | Software › Software engineering | 86.8 | 62.5 | −24.3 |
| SimpleQA Verified (Epoch anti-abstention) | Knowledge & reasoning › General knowledge | 94.0 | 75.2 | −18.8 |
| LMArena Agent Arena Net Improvement (Steering Burden) | Tool & computer use › General agents | 27.4 | 18.4 | −17.5 |
| Harvey LAB (Mercor run, undisclosed ~213-task cohort) | Professional › Legal | 10.0 | 26.6 | +16.6 |
| Humanity's Sixth Sense | Perception & media › Spatial & visual reasoning | 59.5 | 75.9 | +16.4 |
| HLAB-Vals-1.1 | Professional › Legal | 5.0 | 19.6 | +14.6 |
| sudo L7 (October 2026 rescoring) | Software › Software engineering | 52.2 | 38.3 | −13.9 |
| Tax Agent Bench | Professional › Finance, accounting & tax | 67.0 | 79.8 | +12.8 |
| Medical Long Context Reasoning (MLCR-AA, October 2026 methodology) | Health & medicine | 43.9 | 31.7 | −12.2 |
| LMArena Agent Arena Praise vs Complaint | Tool & computer use › General agents | 68.7 | 44.5 | −12.1 |
| AutomationBench 1.0.6 (Zapier) | Tool & computer use › Customer service & operations | 44.3 | 56.3 | +12.0 |
| GDP.xlsx | Professional › General knowledge work | 32.9 | 43.1 | +10.2 |
| LMArena Agent Arena Bash Reliability | Tool & computer use › General agents | 13.1 | -6.0 | −9.5 |
| EnterpriseBench: CoreCraft | Tool & computer use › Customer service & operations | 68.2 | 77.4 | +9.2 |
| Humanity's Last Exam Diamond | Knowledge & reasoning › General knowledge | 80.2 | 71.3 | −8.9 |
| ComplexConstraints (November 2026 rescoring) | Language & writing › Instruction following | 64.5 | 72.9 | +8.4 |
| LMArena Code Arena WebDev | Software › Frontend & design | 1824 | 1764 | −8.0 |
| CritPt v2 (agentic, avg@5) | Science › Physical sciences | 84.9 | 77.4 | −7.4 |
| DAYJOB: Finance (October 2026 rescoring) | Professional › Finance, accounting & tax | 43.7 | 50.5 | +6.8 |
| MysteryMechanism | Math | 51.8 | 45.0 | −6.8 |
| FrontierCode 2.0 (Artificial Analysis) | Software › Software engineering | 49.1 | 43.3 | −5.9 |
| AutomationBench-AA 1.0.6 | Tool & computer use › Customer service & operations | 74.2 | 79.3 | +5.2 |
| Mystery Game Puzzles | Games | 90.0 | 85.0 | −5.0 |
| Bug Hunt Bench | Software › Software engineering | 55.2 | 60.0 | +4.8 |
| Vals Finance Index | Professional › Finance, accounting & tax | 63.6 | 68.3 | +4.7 |
| DAYJOB: Accounting | Professional › Finance, accounting & tax | 45.2 | 49.6 | +4.4 |
| CWE-bench v2 Pass@1 | Software › Cybersecurity | 75.0 | 70.7 | −4.3 |
| Chess Puzzles (task version 2.0) | Games | 79.0 | 83.0 | +4.0 |
| FrontierMath Tiers 1-3 (v2) - Private | Math | 95.1 | 91.2 | −3.9 |
| τ³-Banking (Mercor run, bm25_grep retrieval) | Tool & computer use › Customer service & operations | 55.7 | 59.5 | +3.8 |
| BioMysteryBench (Vals Terminus 2) | Science › Life sciences | 83.7 | 80.0 | −3.7 |
| Vals RSI Index v1.1 (four tasks) | Software › Data & ML | 37.9 | 34.4 | −3.5 |
| Agents' Last Exam v1.1 Score | Professional › General knowledge work | 72.8 | 75.8 | +3.0 |
| DAYJOB: Tax | Professional › Finance, accounting & tax | 15.6 | 12.8 | −2.8 |
| LiveBench 2.0 (2026-11-06 release) | Composite indexes | 84.9 | 87.4 | +2.6 |
| FrontierMath Tier 4 (v2) - Private | Math | 100.0 | 97.6 | −2.4 |
| GPQA Diamond | Science › Multidisciplinary | 93.9 | 96.3 | +2.4 |
| DAYJOB: Healthcare (October 2026 rescoring) | Health & medicine | 45.0 | 47.2 | +2.2 |
| AA-Briefcase v1.1 Elo | Professional › General knowledge work | 1737 | 1699 | −2.1 |
| LMArena Agent Arena Steering Burden | Tool & computer use › General agents | 24.5 | 21.2 | −1.6 |
| LMArena Agent Arena Confirmed Success | Tool & computer use › General agents | 22.5 | 19.2 | −1.6 |
| GDP.pdf (November 2026 rescoring) | Professional › General knowledge work | 51.8 | 53.4 | +1.6 |
| APEX-Agents 1.1 | Professional › General knowledge work | 87.6 | 88.9 | +1.3 |
| Excel Modeling Benchmark v1.1 | Professional › Finance, accounting & tax | 77.4 | 76.5 | −0.9 |
| Finance Agent v2.1 | Professional › Finance, accounting & tax | 69.5 | 68.6 | −0.9 |
| Furniture Assembly Benchmark | Perception & media › Spatial & visual reasoning | 90.0 | 89.2 | −0.8 |
| RiemannBench (November 2026 rescoring) | Math | 83.2 | 84.0 | +0.8 |
| SimpleBench | Knowledge & reasoning › General knowledge | 94.9 | 94.4 | −0.5 |
| APEX-Accounting 1.1 | Professional › Finance, accounting & tax | 70.4 | 70.8 | +0.4 |
| DAYJOB: Law (November 2026 release) | Professional › Legal | 26.5 | 26.2 | −0.3 |
| OTIS Mock AIME 2024-2025 | Math | 99.7 | 100.0 | +0.3 |
| LMArena Agent Arena Tool Hallucination | Tool & computer use › Tool calling | 1.2 | 1.0 | −0.1 |
| Chartography | Perception & media › Documents & charts | 73.4 | 73.5 | +0.1 |
| Vibe Code Bench v1.1 | Software › Frontend & design | 93.8 | 93.7 | −0.1 |
| HANDBOOK.md Agents v1.1 | Tool & computer use › Customer service & operations | 52.3 | 52.3 | 0.0 |
Data: Surge AI, Vals AI, Epoch AI and 14 othersatlas.village.dev