Claude Sonnet 5.5 scores 56 on the Intelligence Index, just 2 points off Opus
As seen on the 24/7 Wall St. homepage on September 28, 2026.
Anthropic's mid-tier model now sits 2 points off its own flagship, and it burns record output tokens to get there, which means more compute billed per task for the cloud providers hosting it.
Anthropic has launched Claude Sonnet 5.5: it scores 56 on the Artificial Analysis Intelligence Index, just 2 points behind Opus 5.5 (max), but at the highest Output Tokens per Task we’ve seen With max effort, Sonnet 5.5 gains 18 points over Sonnet 5 and to #2 on the Intelligence https://t.co/aunF6RTycN
- Replies53
- Reposts41
- Likes658
Continue ReadingShow less
Anthropic's new Claude Sonnet 5.5 lands at 56 on the Artificial Analysis Intelligence Index, second overall and just behind the flagship Opus 5.5. That gap is narrow enough that most real-world tasks will feel indistinguishable from the top tier.
Sonnet 5.5 gains 18 points over Sonnet 5 with max effort applied, a generational leap inside a single product line.
The trade-off is compute. Artificial Analysis notes Sonnet 5.5 produces the highest Output Tokens per Task the firm has ever recorded, which means every call to the model consumes more tokens and therefore more cloud spend than any prior benchmark result.
For businesses running Sonnet at scale, that token volume translates directly into higher API bills, and for the cloud providers hosting the model, it means more compute billed per task. The balance between efficiency and capability has shifted with this release.