Sonnet 5.5
A new version in Anthropic's Claude Sonnet series, positioned within the mid-tier product line of the Claude model family.
Core Facts
Claude Sonnet 5.5 achieved 70.6% on the Terminal benchmark, compared to only 10.3% for the previous generation.
GPT-6.1 Sol and Sonnet 5.5 have the same price per million tokens: $2 for input and $10 for output.
The company states that Sonnet 5 shows significant improvements over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work, with overall performance approaching OPUS 4.8 but at a lower price.
Timeline (last 90 days)
In the Zapier automation benchmark, Argon took 51%, ahead of Sonnet 5.5's 45%.
Opus 5.5 in Max mode scored 58 on the Artificial Analysis Intelligence Index, while Sonnet 5.5 scored 56.
WonderFull's coding Agent score of 68 is on par with ClawedCode (equipped with Sonnet 5.5).
In the Intelligence Index given by the independent evaluation organization Artificial Analysis, Muse Wonder scores only 52, while Claude Opus 5.5 reaches 58.
In test three (generating a 90-day action plan from vague instructions), Opus 5.5 won because it produced a Q4-aligned 90-day sprint plan that was directly executable, while Sonnet generated a local HTML file with dates falling outside Q4.
When a task has a clear specification and verifiable results, a weaker model can deliver results close to a top-tier model as long as it follows the instructions.
In test seven (a YouTube video resource guide, with a mature skill as a fallback), the two versions were extremely similar, and Sonnet 5.5 was twice as fast and $0.4 cheaper, so it won.
In test four (Excel spreadsheet and pitch deck), Opus 5.5 won, as its Excel spreadsheet had formulas in almost every cell that could update automatically, and in this round Opus was faster and cheaper, the only time in the whole session that Opus beat Sonnet on cost-effectiveness.
In test six (scraping that day's AI news from X to build an interactive dashboard), Opus 5.5 was better in a single round of output, but the creator awarded the win to Sonnet 5.5 under the cost rule, believing Sonnet could achieve the same functionality with a few more iterations.
The criterion proposed by the creator: if a task has a clear definition of done, try Sonnet first; if it requires creativity or needs the model to help define "what counts as done," use Opus.
40 more timeline events
All Facts (20)
Claude Sonnet 5.5 achieved 70.6% on the Terminal benchmark, compared to only 10.3% for the previous generation.
75%VerifiedGPT-6.1 Sol and Sonnet 5.5 have the same price per million tokens: $2 for input and $10 for output.
70%VerifiedThe company states that Sonnet 5 shows significant improvements over Sonnet 4.6 in reasoning, tool use, coding, and knowledge work, with overall performance approaching OPUS 4.8 but at a lower price.
70%VerifiedHaiku 5.5 can work as a sub-agent alongside Claude Opus 5.5 and Sonnet 5.5 to complete programming tasks.
65%VerifiedArgon's launch price is $2 per million input tokens and $10 per million output tokens, on par with Sonnet 5.5 and just one-fifth the price of GPT-6 Astra; after the promotional period ends, prices will double to $4 input and $20 output.
65%VerifiedSonnet 5.5's cache read cost remained at $0.20 per million tokens with no reduction, whereas Fable 5.1 cut it by 90% and Opus cut it by approximately 60%.
65%VerifiedIn the deep audit benchmark, Sonnet 5.5's cost is approximately half that of Opus, its score is slightly higher than Opus, and it takes about 5 minutes (roughly half the time Opus takes).
65%VerifiedOn the SWE-Bench Pro benchmark, Sonnet 5 significantly trails Opus 4.8.
65%Partially VerifiedIn the Intelligence Index given by the independent evaluation organization Artificial Analysis, Muse Wonder scores only 52, while Claude Opus 5.5 reaches 58.
70%Partially VerifiedMax mode raises the lower bound of reasoning tokens rather than the upper bound, forcing the model to reach a certain amount of reasoning before it is considered complete, which often harms performance.
65%Partially VerifiedSonnet 5.5 is on par with Opus 5.5 in offensive and defensive cybersecurity capabilities, making it the first model in the Sonnet series to adopt equivalent cybersecurity protection mechanisms.
65%UnverifiedSonnet 5.5 became the first Sonnet model to complete Pokémon Red solely by reading and operating through screenshots.
60%UnverifiedOpus 5.5 in Max mode scored 58 on the Artificial Analysis Intelligence Index, while Sonnet 5.5 scored 56.
50%UnverifiedIn the Zapier automation benchmark, Argon took 51%, ahead of Sonnet 5.5's 45%.
50%UnverifiedWonderFull's coding Agent score of 68 is on par with ClawedCode (equipped with Sonnet 5.5).
50%UnverifiedAnthropic officially recommends: use Sonnet for everyday coding, bug fixing, high-throughput routine development, and polishing documents/slides/spreadsheets; use Opus for tasks requiring complex judgment, long-chain agentic coding, knowledge work, or the toughest problems.
50%UnverifiedWhen a task has a clear specification and verifiable results, a weaker model can deliver results close to a top-tier model as long as it follows the instructions.
50%UnverifiedThe criterion proposed by the creator: if a task has a clear definition of done, try Sonnet first; if it requires creativity or needs the model to help define "what counts as done," use Opus.
50%UnverifiedSonnet 5.5 was the only one of the three models to add background music (rather than just sound effects).
50%UnverifiedAnthropic officially does not recommend using max reasoning mode on Sonnet 5.5, because the model tends to run too long, do work beyond what is needed, and even contradict itself.
50%