Opus 5.5 Beats GPT-6 Sol 6-1 in Seven-Round Showdown

Anthropic's Opus 5.5 and OpenAI's GPT-6 Sol launched on the same day just minutes apart, and blogger @alexverem quickly put them head-to-head across seven representative tasks. Opus 5.5 ultimately won 6:1, with its only loss coming in the lava lamp round, where speed and cost favored Sol. Because the evaluation spans code generation, 3D modeling, hardcore real-world engineering, and subscription value, it has become a widely cited reference for comparing the two models.

Confirmed

• Round 1 (Three.js Eiffel Tower, four-model comparison): Opus 5.5 cost $8.95/10 minutes, GPT-6 Astra $7.45/8 minutes, GPT-6 Sol $3.90/5 minutes, with Kimi K3 also participating; the author found Opus's output the most impressive.
• Round 2 (lava lamp): GPT-6 Sol won in just 50 seconds for 8 cents; Opus 5.5 took 9 minutes 35 seconds and cost $1.27; GPT-6 Astra ran $0.63/5:04, while GPT-6 Luna came in under 1 cent.
• Round 3 (Blender 3D scene from the same prompt): Sol was faster at about $34.55 but had clear rigging and geometry issues; Opus was slower at $48.11, yet the final scene was noticeably more refined in detail and atmosphere.
• Round 4 (single-HTML fishing game): Sol was faster with smaller, cleaner code, but delivered only 1 fish type and 2 upgrades; Opus built a full experience with 5 fish types, rarity tiers, a perfect-catch mechanic, 3 progression trees, dynamic fish schools, a minimap, and more.
• Round 5 (each builds a game from the same prompt): the author called the gap "not even close" and said Opus 5.5 is the best model he has ever seen; this round goes to Opus.
• Round 6 (Intel Arc B70 hands-on): Opus 5.5 got exl3 running in 2 hours and fixed decode, prefill, concurrency, MTP, and vision; the author said he had spent weeks chasing the same goal with Sol without success.
• Round 7 (subscription value): under Claude Pro ($20), Opus 5.5 on high ran for nearly 2 hours straight before hitting the 5-hour cap, with each 5-hour window accounting for only about 15% of weekly usage — far more, in the author's view, than what $20 ChatGPT Plus offers.
• Final score: Opus wins 6 rounds (Eiffel Tower, Blender, fishing game, same-prompt game, GPU hands-on, subscription quota); Sol wins 1 (lava lamp).

Unconfirmed

• All conclusions come from a single evaluation by @alexverem, who self-describes as partial to Opus, so the results carry subjective judgment and have not been independently reproduced.

Why it matters

• Two flagship models launching on the same day is rare in itself, and this seven-round showdown covers creative generation, engineering quality, and cost efficiency, offering a multi-dimensional reference for model selection. In particular, the hands-on test on a long-horizon hardcore task like the Intel Arc B70 and the real-world $20 subscription quota testing bear directly on the actual costs and productivity decisions of heavy users.

2026-09-25 ~ 2026-09-25 · 9 related posts

• [source] Opus 5.5 and GPT-6 Sol launched minutes apart; author scores 7 matchups — alex_verem · 2026-09-25
• Three.js Eiffel Tower, four models: Opus 5.5 shines, Kimi K3 best value — alex_verem · 2026-09-25
• Lava lamp test: GPT-6 Sol wins on cost and speed at $0.08 in 50 seconds — alex_verem · 2026-09-25
• Blender head-to-head: Sol saves ~$14, Opus delivers the polished scene — alex_verem · 2026-09-25
• Fishing game test: Sol optimizes the benchmark, Opus builds the product — alex_verem · 2026-09-25
• Same-prompt game build: Opus 5.5 vs GPT-6 Sol "not even close" — alex_verem · 2026-09-25
• [source] Opus 5.5 gets exl3 running on Intel Arc B70 GPUs in 2 hours — alex_verem · 2026-09-25
• $20 Claude Pro runs Opus 5.5 high for nearly 2 hours per 5-hour window — alex_verem · 2026-09-25
• [source] Final tally: Opus 5.5 dominates GPT-6 Sol 6-1 across seven tests — alex_verem · 2026-09-25