Live arena ratings and prices, updated daily. The most-searched model question, answered with data.
Live data · updated 2026-09-15 18:18 · sources: LMArena + OpenRouter
Right now: Claude leads on quality (1506 vs 1257 arena). claude-sonnet-4-6-search is the smarter buy when budget matters.
| Dimension | Claude (claude-fable-5) | GPT (gpt-5.6-sol-xhigh) | Winner |
|---|---|---|---|
| Arena rating (live) | 1506 (claude-fable-5) | 1257 (gpt-5.6-sol-xhigh) | Claude |
| Models in top 10 | 10 | 9 | Claude |
| Context window | 1M | 1.05M | GPT |
| Flagship output $/1M | $50 | $20 | GPT |
| Best value pick | claude-sonnet-4-6-search ($7.5) | gpt-5.6-sol-xhigh (codex-harness) ($20) | Claude |
Code quality is your top priority — Anthropic currently holds 10 of the top-10 arena spots, and the Claude family is consistently strong on frontend/UI generation.
You want the broader ecosystem and a lower flagship price — GPT's best-value option (gpt-5.6-sol-xhigh (codex-harness)) undercuts Claude's flagship on output price while staying near the top of the board.