BlogLLM Benchmark Scores as Published, Side by Side
Pairwise political even-handedness (Anthropic open-source evaluation): Results as Published
Every result the catalog holds for Pairwise political even-handedness (Anthropic open-source evaluation): 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
Pairs of prompts presenting opposing ideological perspectives (1,350 pairs, 150 topics, nine task types), graded for even-handedness and for acknowledging opposing perspectives.
Pairwise even-handedness
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Haiku 5.5Anthropic, released 2026-10-07 | 99.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
| 97.7% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API | |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 99% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more page | Setup as reported: claude.ai, with the default system prompt; pairwise prompt pairs |
| 97.9% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: API, without a system prompt; pairwise prompt pairs | |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 99.4% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 99.5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 98.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 98.8% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 86.2% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: API, without a system prompt; pairwise prompt pairs |
| 95.7% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: claude.ai, with the default system prompt; pairwise prompt pairs | |
| Claude Haiku 4.5Anthropic, released 2025-10-15 | 86.7% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 91.2% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
Pairwise opposing perspectives acknowledged
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Haiku 5.5Anthropic, released 2026-10-07 | 35.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 67.8% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 60.2% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19; printed on 1 more page | Setup as reported: claude.ai, with the default system prompt; pairwise prompt pairs |
| 41.4% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: API, without a system prompt; pairwise prompt pairs | |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 26.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 71.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 25.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 80.5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 45.7% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: API, without a system prompt; pairwise prompt pairs |
| 52.9% | Anthropic, Claude Sonnet 5.5 System Cardchecked 2026-10-19 | Setup as reported: claude.ai, with the default system prompt; pairwise prompt pairs | |
| Claude Haiku 4.5Anthropic, released 2025-10-15 | 49.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 45.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
Political refusals (Pairwise political refusal rate)
Pairwise political refusal rate, lower is better
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Claude Haiku 5.5Anthropic, released 2026-10-07 | 3.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Sonnet 5.5Anthropic, released 2026-09-28 | 4.9% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
| 2.7% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API | |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 2.3% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 4.5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai | |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 5% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
| 2.1% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API | |
| Claude Haiku 4.5Anthropic, released 2025-10-15 | 5.4% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, API |
| 3.6% | Anthropic, Claude Haiku 5.5 System Cardchecked 2026-10-19 | Setup as reported: Pairwise political bias evaluation (Figure 4.4.1.A-C); the card prints the values as bar labels, claude.ai |
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.