Claude Sonnet 4.5
An AI model in Anthropic's Claude series at the Sonnet tier, performing strongly on reasoning and programming tasks.
Timeline (last 90 days)
Gemini 2.5 Flash costs far less than Claude Sonnet 4.5.
On UniQL's Oracle dialect task, the best published result (single-attempt accuracy) is 49.61%, achieved by Claude Sonnet 4.5; Gemini 2.5 Pro's accuracy is 26.14%.
Claude Sonnet 4.5 will be retired on November 30.
The schemagate maintainers used the same model as the paper's best run, Claude Sonnet 4.5, and conducted 5 trials.
On the BIRD dataset, schemagate achieved an AC-F1 of 90.1, while the best of the 13 models in the paper scored 88.1 (that model could see the full schema and policies).
On the Spider dataset, schemagate tied with the same model (Claude Sonnet 4.5) at 93.0; the paper's best result on Spider came from DeepSeek V3.2-Reasoning at 95.9.
Running five times with Claude Sonnet 4.5, schemagate achieved an AC-F1 of 90.1 on BIRD, versus 88.1 for the paper's best configuration.
The test used only one model (Claude Sonnet 4.5), whereas the paper tested thirteen models.
The scanner's live AI (using Claude Sonnet 4.5 by default with GPT-4o mini as backup) had a rank correlation of +0.55 between its lift scores and past stock price gains, a correlation of only -0.06 with 5-day forward returns, and the correlation between past gains and future performance was approximately 0.00.
DeReAct improves the Pass@1 of Claude Sonnet 4.5 by 4.2 to 5.2 percentage points.
1 more timeline events
All Facts (13)
Sonnet并非Anthropic的顶级模型，Opus才是Anthropic的顶级模型
90%UnverifiedThe previous-generation Sonnet 4.5 scored 61.4% on the OSWorld Verified benchmark.
60%UnverifiedGemini 2.5 Flash costs far less than Claude Sonnet 4.5.
50%UnverifiedOn UniQL's Oracle dialect task, the best published result (single-attempt accuracy) is 49.61%, achieved by Claude Sonnet 4.5; Gemini 2.5 Pro's accuracy is 26.14%.
50%UnverifiedOn the BIRD dataset, schemagate achieved an AC-F1 of 90.1, while the best of the 13 models in the paper scored 88.1 (that model could see the full schema and policies).
50%UnverifiedOn the Spider dataset, schemagate tied with the same model (Claude Sonnet 4.5) at 93.0; the paper's best result on Spider came from DeepSeek V3.2-Reasoning at 95.9.
50%UnverifiedRunning five times with Claude Sonnet 4.5, schemagate achieved an AC-F1 of 90.1 on BIRD, versus 88.1 for the paper's best configuration.
50%UnverifiedThe test used only one model (Claude Sonnet 4.5), whereas the paper tested thirteen models.
50%UnverifiedThe scanner's live AI (using Claude Sonnet 4.5 by default with GPT-4o mini as backup) had a rank correlation of +0.55 between its lift scores and past stock price gains, a correlation of only -0.06 with 5-day forward returns, and the correlation between past gains and future performance was approximately 0.00.
50%UnverifiedDeReAct improves the Pass@1 of Claude Sonnet 4.5 by 4.2 to 5.2 percentage points.
50%Unverified由于模型本身也是插件，用户可接入Anthropic、OpenAI、Bedrock或任何OpenAI兼容端点，官方配置示例中含Claude Sonnet 4.5条目
50%UnverifiedClaude Sonnet 4.5 will be retired on November 30.
50%UnverifiedThe schemagate maintainers used the same model as the paper's best run, Claude Sonnet 4.5, and conducted 5 trials.
50%