{"slug": "gemini-4-neon-looks-like-an-opus-5-5-competitor", "title": "Gemini 4 Neon looks like an Opus 5.5 competitor", "summary": "Google released Gemini 4 Neon, which the Agent Duel leaderboard analysis projects as roughly equivalent to Opus 5.5 in performance while serving at 127 tokens per second on OpenRouter versus 46 tokens per second for Gemini 4 Carbon. The projection is based on what a faster Gemini 4 Carbon might score, since Carbon ranks as the best model on the Agent Duel leaderboard when early votes are excluded, with its main drawback being speed.", "body_md": "*TLDR: we project Gemini 4 Neon’s performance based on what a faster Gemini 4 Carbon might score and find it equivalent to Opus 5.5*\n\nGemini 4 Carbon is the best model on the [Agent Duel leaderboard](https://agentduel.com/leaderboard?excludeEarlyVotes=true) if you exclude early voting. Carbon’s main problem is that it is [slooooow](https://micjm.com/blog/carbon-is-that-slow/).\n\nToday, [Gemini 4 Neon was released](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-neon/), and looks to be an incremental benchmark improvement with a very large speed improvement. Earlier today, it was being served extremely fast via OpenRouter: [127 tokens per second](https://openrouter.ai/google/gemini-4-neon). This is way faster than [Carbon at 46 tokens per second](https://openrouter.ai/google/gemini-4-carbon).", "url": "https://wpnews.pro/news/gemini-4-neon-looks-like-an-opus-5-5-competitor", "canonical_source": "http://micjm.com/blog/gemini-4-neon-looks-like-a-top-model/", "published_at": "2026-11-18 06:00:00+00:00", "updated_at": "2026-11-19 00:42:26.370958+00:00", "lang": "en", "topics": ["large-language-models", "ai-research", "ai-products"], "entities": ["Google", "Gemini 4 Neon", "Gemini 4 Carbon", "Opus 5.5", "Agent Duel", "OpenRouter"], "also_reported_by": [], "takeaway": "Gemini 4 Neon serves at 127 tokens per second on OpenRouter versus 46 tokens per second for Gemini 4 Carbon, and is projected to match Opus 5.5 in benchmark performance.", "recommendation": {"advice": "When latency matters, evaluate Gemini 4 Neon over Gemini 4 Carbon: it served at 127 tokens per second on OpenRouter versus Carbon's 46, with projected Opus 5.5-level scores.", "playbook": "inference-serving", "playbook_url": "https://wpnews.pro/playbooks/inference-serving"}, "alternates": {"html": "https://wpnews.pro/news/gemini-4-neon-looks-like-an-opus-5-5-competitor", "markdown": "https://wpnews.pro/news/gemini-4-neon-looks-like-an-opus-5-5-competitor.md", "text": "https://wpnews.pro/news/gemini-4-neon-looks-like-an-opus-5-5-competitor.txt", "jsonld": "https://wpnews.pro/news/gemini-4-neon-looks-like-an-opus-5-5-competitor.jsonld"}}