Opus 5.5 sweeps GPT-6 Sol on every benchmark, but Sol runs tasks for $1.06
Vast-Grapefruits · reddit · 2026-09-23
A Redditor compiled a head-to-head of Anthropic's Opus 5.5 and OpenAI's GPT-6 Sol, which shipped the same day.
- Raw scores: Opus 5.5 wins every benchmark in the table.
- Surprise: On GDPval, the "real office work" eval, Sol scores roughly 100 Elo below the GPT-5.6 Sol it replaces — Artificial Analysis attributes the drop to weaker presentation and deliverables skipping required task parts, not reasoning failures.
- Cost edge: Sol costs half per token, and at max effort Opus 5.5 emits 119K output tokens per task vs 31K for Sol, which runs the full index for about $1.06 per task.
The author is waiting for real-world feedback over the coming days.
Related event: Anthropic Launches Claude Opus 5.5: Stronger, 40% Cheaper, Tops AA Index(259 posts)→
More from Models
- Robot build test: GPT-6 Astra stabilizes two-board structure, Claude Fable 5.2 falls short — qinzytech · 2026-11-12
- Zvi dissects Anthropic's latest system card, disputing the autonomy threshold call — Don't Worry About the Vase (Zvi) · 2026-11-12
- Unsloth ships Qwen3.8-27B Dynamic V3.1 GGUFs with 5-10% accuracy gains — danielhanchen · 2026-11-12
- antirez backs low-bit DeepSeek 4.1 Flash over Qwen 27B dense on Apple M5 Macs — antirez · 2026-11-12
- Gemini re-renders images as text when vision is removed, dropping only 6% WSR — zacharynado · 2026-11-12
- Bare-harness coding benchmark: Fable 5.2 best overall, Gemini 4 Carbon best value — Semantic_meaning · 2026-11-12