PINNACLE Agentic Benchmark: Qwen3.8 Makes 15% Fewer Errors Than Claude Opus 5
ryanshrout · x · 2026-09-15
Signal65 scored eight new model configurations on its PINNACLE agentic benchmark, five of them open weights from Qwen, DeepSeek and Zai.
Key finding: Qwen3.8 made 15% fewer errors than Claude Opus 5 in their testing, and its 180B sibling runs on a desktop. The authors argue open weights are closer to the hosted frontier than the pacing debate admits — meaning any agreement to slow frontier development favors whoever can run their own hardware. Three labs atop the board (OpenAI, xAI, Anthropic) just backed a call to pace frontier AI.
More from Models
- Claude upgrade-page traffic rebounded to 17.6M in September, near May's 18.6M peak — firstadopter · 2026-11-02
- Open-source GLiNER-PII scrubs 13 identifiers from a discharge note in 550ms on GPU — MaziyarPanahi · 2026-11-02
- OpenAI Dots user says bloat and broken infra drove him to try newcomer Muse — Koala_Confused · 2026-11-02
- Fable 5.5 reportedly one-shots animation in pure code as Anthropic's coding bet pays off — Hesamation · 2026-11-02
- One agent goal has been running for 4 days, burning 30K tokens — CtrlAltDwayne · 2026-11-02
- Daily AI Brief: Gemini 4 Carbon spotted on Arena, DeepSeek open-sources coding agent — testingcatalog · 2026-11-02