You’re reading the October 2026 report. See November 2026, the latest →
Labs
Teams pick a lab, then move between its models. Each lab page checks every model the Benchmark tests from that lab against the same line.
Anthropic 2 models: Claude Opus 5.5, Claude Sonnet 5.5 DeepSeek 1 model: DeepSeek V4.1 Flash Google 3 models: Gemini 3.8 Flash, Gemma 4 31B, Gemini 4 Argon Meta 1 model: Muse Wonder Moonshot AI 1 model: Kimi K3 NVIDIA 1 model: Nemotron 3.5 Super VL OpenAI 4 models: GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna, GPT-5.6 Luna xAI 1 model: Grok 4.7 Z.ai 1 model: GLM 5.3