Terminal-Bench 2.1 ↗87.64 %
Claude Opus 5.5 (Terminus 2) · Terminal-Bench 2.1 · Score ↗
Anthropic / MODEL EVIDENCE
Claude Opus 5.5 (Terminus 2)
1 benchmarks · 1 observations · 1 sources
Scores at a glance
1 benchmarks · large dot: Claude Opus 5.5 (Terminus 2); small dots: closest models where reported
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
1 of 1 shared benchmark comparisons
1 of 1 shared benchmark comparisons
Claude Fable 5.1 (Terminus 2) →Anthropic
1 of 1 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
POST-TRAINING / SOURCE DISCLOSURE
What the Claude Opus 5.5 report says
Anthropic describes fine-tuning after pretraining to align the assistant with Claude’s constitution. The system card also documents human preference selection, safety evaluation and adversarial testing.
- Post-training and fine-tuning guided by Claude’s constitution
- Human preference and safety-evaluation data
The system card does not disclose a production PPO, DPO or GRPO recipe. Constitutional AI is a related method note; its particular SL-CAI/RL-CAI implementation is not asserted for this release.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
Names used in source tables
Claude Opus 5.5 (Terminus 2)