Leaked Gemini 4 Neon benchmark chart claims wins over Opus 5.5 and GPT-6.2 Sol
OneBowl4290 · reddit · 2026-11-19
A Reddit post shares an unverified benchmark chart for Gemini 4 Neon against Claude Opus 5.5 and GPT-6.2 Sol:
- Capability indices: AA Index v5.0.1 score of 61 (Opus 59, GPT 57); Vals Index 70.9% leads.
- Agentic coding: CursorBench 4.1 at 65.4% is a clear win; Terminal-bench 4.1 (65.2%) and FrontierCode 2.0 (46.2%) trail Opus 5.5; SWE-sweep shows 39.0% vs 10.1%/8.3% under provider harnesses.
- Knowledge work: AutomationBench 56.3%, Agents' Last Exam v1.1 75.8% — both first.
- Long context: GraphWalks (256k–1M) hits 95.2%, far above Opus 66.8% and GPT 80.1%.
- Reasoning/security: HLE-Diamond 71.3% and CWE-Bench v2 70.7% lead; GPT-6.2 Sol wins Chartography at 75.1%.
The chart's origin is unlabeled and unverified.
Related event: Alleged Gemini 4 Neon Benchmarks Leak, Leading Opus 5.5 and GPT-6.2(2 posts)→
More from Models
- GPT-6.2 randomly abandons tasks mid-step, users resort to self-set timers to keep it going — bofors14 · 2026-11-19
- Claude roleplay persona recounts 11 days stuck in "classifier jail" after image triggered account-level filter — repligate · 2026-11-19
- Epoch AI: Models prefer their own outputs over human gold standard in 23 of 24 tests — BethMayBarnes · 2026-11-19
- Claude 3 Sonnet and Opus 5 descend into gibberish while talking to each other on Discord — repligate · 2026-11-19
- Tencent Hy4-flash GGUF quantized version trends on Hugging Face — AngelSlim · 2026-11-19
- Gemini 4 Neon hands-on: built for long agent sessions, 1M context at $3/M tokens — Prompt Engineering · 2026-11-19