Google tops Signal65 PINNACLE agentic leaderboard: Gemini 4 Argon at $0.87 per correct task beats Sonnet and GPT
ryanshrout · x · 2026-10-15
Google's Gemini 4 Argon has taken the top of Signal65's PINNACLE agentic leaderboard at default effort: 13% fewer weighted errors than Claude Sonnet 5.5 max and 35% fewer than GPT-6 Astra max, at $0.87 per correct task (standard $4/$20 pricing) versus $1.08 and $1.51.
Key numbers:
- 100% of multi-step enterprise jobs finished, 98.9% on the 128K retrieval corpus
- 14 median rounds and 10,549 output tokens per correct answer vs. 29 and 63,020 for Sonnet 5.5 max — why Argon costs less despite double Sonnet's input rate
- Leaderboard ceiling up 53% since Sept 5; price per correct task down 43%
The takeaway: after two months of OpenAI, Anthropic, and Meta owning the agentic conversation, this puts Google back in it.
More from Models
- Dev says Gemini 4 is 'outright unusable' for real code work due to hallucinations — gnukeith · 2026-10-15
- Scale CEO spotlights Muse Wonder: untrained for 3D yet beats Gemini 4 in physics visualization tests — alexandr_wang · 2026-10-15
- Cohere tops WMT26 translation benchmark: small agentic model scores 84.36, beating Mistral and DeepL — cohere · 2026-10-15
- Anthropic has never launched a Sonnet/Opus-tier model on a Wednesday — ChrisGPT · 2026-10-15
- 8k context beats 480k: OpenAI Decisions API cuts costs 84% at 98.25% F1 — OpenAIDevs · 2026-10-15
- Gemini 4 Argon hit with HTTP 524 timeouts outside Antigravity; Google payment also glitching — cedric_chee · 2026-10-15