Claude Opus 4.7 Arrives: Reclaims the Throne with 87.6% on SWE-bench, Leaving GPT-5.4 and Gemini Behind

product#llm📝 Blog|Analyzed: Apr 20, 2026 00:26
Published: Apr 20, 2026 00:23
1 min read
Qiita AI

Analysis

Anthropic has unleashed a monumental upgrade with Claude Opus 4.7, dramatically improving performance while keeping prices exactly the same. This Large Language Model (LLM) doesn't just edge out the competition; it completely leaves GPT-5.4 and Gemini 3.1 Pro in the dust on coding and agent benchmarks. It's a thrilling development for developers and tech leads looking for top-tier AI power without a price hike.
Reference / Citation
View Original
"Opus 4.7 is a new supercar with the engine swapped out while keeping the same body. The price remains unchanged ($5 / $25 per 1M tokens), and the API specs are mostly the same. Yet the 0-100 acceleration (SWE-bench Verified) jumped from 80.8% → 87.6%."
Q
Qiita AIApr 20, 2026 00:23
* Cited for critical analysis under Article 32.