Gemini 4 Neon benchmark tables leak showing leads over Opus 5.5 and GPT-6.2 Sol
After the launch of Gemini 4 Neon, its official benchmark comparison chart has been circulating widely on X. Several posters, including Saboo Shubham, imjustnewatai, OkBarracuda1161, and YashasGunderia, shared the same table, describing Neon as combining frontier-level reasoning with fast task completion, aimed at agent workloads, and inviting skeptics to test it themselves.
Confirmed
- The comparison targets are Claude Opus 5.5 and GPT-6.2 Sol, with the chart dated November 2026 and covering roughly 20 benchmarks.
- Specific data points include: 61 for Neon on AA Index v5.0.1 (vs. 59/57 for rivals), 95.2% on GraphWalks long-context, and 89.0% on LVBench multimodal.
- imjustnewatai's tally shows Neon beating both Sol 6.2 and Opus 5.5 on 17/20 benchmarks; against the flagship Astra 6.2 and Fable 5.2, it wins on both in 4/12 benchmarks.
- imjustnewatai also emphasized that Neon is priced at one-third of its rivals.
- Posts disagree on the benchmark count (15/16/18/20), but all point to the same comparison table.
Unconfirmed
- The data is labeled as an "official" benchmark table, but everything came via poster reposts; no first-party Google publication appeared in this batch of posts.
Why it matters
If the numbers hold up, Neon outperforming both rivals' flagships on most benchmarks at about one-third the price would directly disrupt model selection and pricing for agent workloads—which also explains the table's viral spread in the community.
2026-11-19 ~ 2026-11-19 · 5 related posts
Primary sources
- Gemini 4 Neon benchmark table shows wins over Claude Opus 5.5 and GPT-6.2 Sol on most of 20 evals — Saboo_Shubham_ ·
- Gemini 4 Neon benchmark table shows wins over Opus 5.5 and GPT-6.2 Sol on most metrics — imjustnewatai ·
- Gemini 4 Neon Benchmark Sheet: Beats Opus 5.5 and Sol 6.2 on 17/20, at One-Third the Cost — imjustnewatai ·
- Gemini 4 Neon official benchmarks: leads Opus 5.5 on most of 16 evals — Ok_Barracuda_1161 · 2026-11-19
- [source] Gemini 4 Neon benchmark table shows wins over Claude Opus 5.5 and GPT-6.2 Sol on most of 20 evals — Saboo_Shubham_ · 2026-11-19
- [source] Gemini 4 Neon Benchmark Sheet: Beats Opus 5.5 and Sol 6.2 on 17/20, at One-Third the Cost — imjustnewatai · 2026-11-19
2 near-duplicate retellings: imjustnewatai · YashasGunderia