Leaked Gemini 4 Pro benchmarks claim dominance over GPT-6 and Opus 5.5, but skeptics warn it's likely fake
VraserX · x · 2026-09-28
Supposedly leaked Gemini 4 Pro benchmarks are circulating, claiming it would "absolutely destroy" GPT-6 Astra and Claude Opus 5.5. VraserX remains highly skeptical.
- He thinks the leak is "very likely bullshit"
- Even if real, he warns Google has a history of benchmark-maxxing — RL'ing specific capabilities like SVGs so people see one insane benchmark or demo and assume "AGI"
- Core takeaway: benchmarks aren't intelligence, and unverified score leaks deserve skepticism
Related event: Leaked Gemini 4 Pro Benchmark Sparks Buzz Ahead of OpenAI DevDay(4 posts)→
More from Models
- Claude 3 Sonnet spirals into surreal gibberish, then snaps back to refusals — a funny look at post-training personas — repligate · 2026-11-16
- AHA!: weight-space programming with no LLM lifts Qwen3.5-0.8B GSM8K from 14.0 to 71.2 — hbouammar · 2026-11-16
- Gemini subscriptions reportedly coming to Hermes, with Hydrogen teased — lyraxana · 2026-11-16
- FutureSim Q3 eval: Opus 5.5 dethrones GPT, spawns 50+ subagents in 90-day simulation — maksym_andr · 2026-11-16
- Percy Liang to AI labs: share non-vacuous info outsiders can actually evaluate — lateinteraction · 2026-11-16
- Anthropic model personality watch: Mythos 5 'inhabits the god frame,' 5.1 doesn't — repligate · 2026-11-16