Artificial Analysis ranks GPT-6 and Claude models on FrontierCode cost and time

Artificial Analysis says four models sit on the Pareto frontier of score versus cost per task on FrontierCode: GPT-6 Luna (max) at 42% and $0.09 per task, MiMo-V2.6-Pro at 45% and $0.18, GPT-6.1 Sol (max) at 49% and $0.81, and Claude Opus 5.5 (max with fallback) at 52% and $10.15.

Cost per task across the benchmark ranges from $0.09 to $12.00 for Claude Fable 5.1 (max with fallback). The firm also says GPT-6 models led on token efficiency in its testing.

A separate comparison by time per task puts GPT-6 Luna (max), GPT-6 Astra (max), GPT-6.1 Sol (max) and Claude Opus 5.5 (max with fallback) on the score-versus-time frontier. Artificial Analysis estimates time from each model's token usage and its live output speed measurements. It says GPT-6 models finish quickly on 27k to 48k output tokens per task. DeepSeek V4.1 Flash (max) uses about 119k tokens but averages roughly 7 minutes per task, around 40% of Claude Opus 5.5's 17 minutes.