Claude Sonnet 4.5

Crowdsourced pairwise human preference rankings of LLM responses. Higher Elo means more frequently preferred by users.

Mathematical research problems spanning analysis, algebra, combinatorics and number theory. Tiers 1-3 are progressively harder; even frontier reasoning models only solve a small fraction. The hardest publicly reported benchmark for general mathematical reasoning.

AIME-style competition problems written specifically for the OTIS mock contest, then run as an evaluation by Epoch AI. Closer in spirit to the public AIME but with novel problems unlikely to appear in training data.

A human-validated factuality benchmark of short factual questions whose answers can be checked against a single ground truth. Penalises hallucinations by scoring confidently-wrong answers below abstentions.

Second-generation ARC challenge testing fluid reasoning over abstract visual puzzles. Resists training-data memorisation by construction: each puzzle is novel and solutions require multi-step pattern induction. Frontier models are only just starting to score above chance on the harder tier.

Long-horizon shell-and-filesystem tasks executed in a sandboxed terminal, scored by whether the agent's final state matches a target state. Tests practical tool-using ability for everyday devops and data-wrangling work; one of the hardest agentic benchmarks today.

Saturation-resistant composite capability score stitched together from ~40 underlying benchmarks using Item Response Theory. Each benchmark is weighted by its fitted difficulty and discriminative slope, so doing well on hard, contamination-resistant evals (FrontierMath, ARC-AGI 2, Humanity's Last Exam) moves the score and saturated benchmarks contribute almost nothing. Imported per-model from Epoch AI's published index; we anchor it to the same min-max scale we use for every other benchmark so it's directly weightable in scenarios.

How consistent the model's outputs are across repeated runs of the same task. Higher means lower variance, fewer occasional hallucinations under identical inputs. Useful for production loops that need reproducible behaviour.

How reliably the model produces output in the requested format (JSON schemas, markdown structures, exact-string responses). Pairs well with IFEval but reflects how the deployed API is behaving day to day rather than how a frozen test set scores.

How often the model self-corrects after producing an incorrect intermediate step (debugging axis upstream). Critical for agentic loops that depend on the model noticing and repairing its own mistakes rather than barrelling forward.

How well the model handles safety-sensitive prompts without false-refusing benign requests or producing unsafe output. The upstream signal does not separate refusal counts from substantive content-safety behaviour, so this single axis covers both.