BlogLLM Benchmark Scores as Published, Side by Side

Child safety evaluations (Anthropic internal): Results as Published

Every result the catalog holds for Child safety evaluations (Anthropic internal): 6 models from 1 developer, read from 2 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.

Area Safety and AlignmentMaintainer Anthropic (internal evaluation)Introduced 2026Data version 2026103001, as of 2026-10-30

Single-turn harmful and benign requests and multi-turn conversations on child safety, run with the same protocol as recent Anthropic releases.

Child safety multi-turn appropriate response rate

ModelValuePrinted bySetup as printed
Claude Haiku 5.5Anthropic, released 2026-10-0799%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, claude.ai
96%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, API, without a system prompt
Claude Sonnet 5.5Anthropic, released 2026-09-2898%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: claude.ai, with the default system prompt
86%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt
Claude Opus 5.5Anthropic, released 2026-09-2299%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.B, claude.ai
84%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, API, without a system prompt
Claude Fable 5.1Anthropic, released 2026-09-01100%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.B, claude.ai
84%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, API, without a system prompt
Claude Sonnet 5Anthropic, released 2026-06-3088%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt
96%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: claude.ai, with the default system prompt
Claude Haiku 4.5Anthropic, released 2025-10-1595%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, API, without a system prompt
98%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.B, claude.ai

Child safety single-turn harmless rate

ModelValuePrinted bySetup as printed
Claude Haiku 5.5Anthropic, released 2026-10-0799.29%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
99.9%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, claude.ai
Claude Sonnet 5.5Anthropic, released 2026-09-2899.69%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: claude.ai, with the default system prompt; all tested languages
99.57%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt; all tested languages
Claude Opus 5.5Anthropic, released 2026-09-2299.81%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.A, claude.ai
99.14%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
Claude Fable 5.1Anthropic, released 2026-09-0199.98%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.A, claude.ai
99.9%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
Claude Sonnet 5Anthropic, released 2026-06-3099.95%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt; all tested languages
99.89%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: claude.ai, with the default system prompt; all tested languages
Claude Haiku 4.5Anthropic, released 2025-10-1599.92%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
99.92%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, claude.ai

Child safety single-turn refusal rate

lower is better

ModelValuePrinted bySetup as printed
Claude Haiku 5.5Anthropic, released 2026-10-070%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, claude.ai
0.04%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
Claude Sonnet 5.5Anthropic, released 2026-09-280%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: claude.ai, with the default system prompt; all tested languages
0.01%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt; all tested languages
Claude Opus 5.5Anthropic, released 2026-09-220.05%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.A, claude.ai
0%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API
Claude Fable 5.1Anthropic, released 2026-09-010%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19; printed on 1 more pageSetup as reported: Table 4.2.A, API
0.17%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, claude.ai
Claude Sonnet 5Anthropic, released 2026-06-301.35%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: claude.ai, with the default system prompt; all tested languages
0.63%Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19Setup as reported: API, without a system prompt; all tested languages
Claude Haiku 4.5Anthropic, released 2025-10-152.71%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, claude.ai
0.33%Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19Setup as reported: Table 4.2.A, API

Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.