Daniel
@0shotdev.bsky.social
Software Engineer | MERN x GenAI | Writing about LLMs -> what's hype, what's real.
2/ The highlights:
• Effort parameter (beta, API): Opus 4.5 medium > Sonnet 4.5 quality, 76% fewer tokens
• Tool Search (beta): ~77K → ~8.7K context tokens
• Programmatic Tool Calling (beta): process data outside context
• 1.5x more prompt injection resistant
• Pricing: $5/$25 per MTok (3x cheaper)
• Effort parameter (beta, API): Opus 4.5 medium > Sonnet 4.5 quality, 76% fewer tokens
• Tool Search (beta): ~77K → ~8.7K context tokens
• Programmatic Tool Calling (beta): process data outside context
• 1.5x more prompt injection resistant
• Pricing: $5/$25 per MTok (3x cheaper)
November 25, 2025 at 6:02 PM
2/ The highlights:
• Effort parameter (beta, API): Opus 4.5 medium > Sonnet 4.5 quality, 76% fewer tokens
• Tool Search (beta): ~77K → ~8.7K context tokens
• Programmatic Tool Calling (beta): process data outside context
• 1.5x more prompt injection resistant
• Pricing: $5/$25 per MTok (3x cheaper)
• Effort parameter (beta, API): Opus 4.5 medium > Sonnet 4.5 quality, 76% fewer tokens
• Tool Search (beta): ~77K → ~8.7K context tokens
• Programmatic Tool Calling (beta): process data outside context
• 1.5x more prompt injection resistant
• Pricing: $5/$25 per MTok (3x cheaper)
1/
The loop continues - Claude Opus 4.5 dropped.
First model to break 80% on real-world software engineering (SWE-bench Verified).
But the interesting part isn't just the benchmark, it's also what Anthropic is doing to make their smartest model usable day-to-day.
#Claude #Anthropic #Opus #GenAI
The loop continues - Claude Opus 4.5 dropped.
First model to break 80% on real-world software engineering (SWE-bench Verified).
But the interesting part isn't just the benchmark, it's also what Anthropic is doing to make their smartest model usable day-to-day.
#Claude #Anthropic #Opus #GenAI
November 25, 2025 at 6:01 PM
1/
The loop continues - Claude Opus 4.5 dropped.
First model to break 80% on real-world software engineering (SWE-bench Verified).
But the interesting part isn't just the benchmark, it's also what Anthropic is doing to make their smartest model usable day-to-day.
#Claude #Anthropic #Opus #GenAI
The loop continues - Claude Opus 4.5 dropped.
First model to break 80% on real-world software engineering (SWE-bench Verified).
But the interesting part isn't just the benchmark, it's also what Anthropic is doing to make their smartest model usable day-to-day.
#Claude #Anthropic #Opus #GenAI