Arena.ai ranks GPT-6 Sol (Max) fourth in Code Arena WebDev

Arena.ai reports that GPT-6 Sol (Max) scored 1,689 points and ranked fourth in Code Arena: WebDev. Arena.ai says the result reshapes the Pareto frontier at $8 per million tokens for blended input and output. It describes that performance as on par with Claude Opus 5 (Max) at less than half the cost, and two Pareto positions behind the frontier leader, GPT-6 Astra (Max), at one-fifth the cost.

Arena.ai says GPT-6 Sol (Max) is 72 points higher than GPT-5.6 Sol (xHigh), which ranked 17th. It lists these category rank changes: Simulations from 18th to 3rd, Content Creation from 14th to 3rd, Gaming from 13th to 3rd, Consumer Product from 14th to 4th, reference-based design from 10th to 4th, Brand & Marketing from 13th to 5th, and Data & Analytics from 14th to 6th. Arena.ai says GPT-6 Luna’s score is not yet available because real-world votes are still coming in.

OpenAI introduced GPT-6 Sol and GPT-6 Luna as faster, more affordable models that bring much of GPT-6 Astra’s strengths to work at scale. OpenAI says more efficient caching and inference allow it to set API prices for Sol and Luna 50% lower than GPT-5.6 promotional pricing.