ARC AI2 ↗86.3%
Claude Instant · ARC AI2 · Challenge score · % ↗
1 more reported result
85.7% · Claude Instant · ARC AI2 · Challenge score · % · Epoch AI · ARC AI2 ↗
Anthropic / MODEL EVIDENCE
Claude Instant
6 benchmarks · 13 observations · 5 sources · released 2023-08-09
Scores at a glance
6 benchmarks · large dot: Claude Instant; small dots: closest models where reported
GSM8K ↗86.7%
Claude Instant · GSM8K · EM · % ↗
1 more reported result
80.9% · Claude Instant · GSM8K · EM · % · Epoch AI · GSM8K ↗
MMLU ↗68.8%
Claude Instant · MMLU · EM · % ↗
3 more reported results
73.4% · Claude Instant · MMLU · EM · % · Epoch AI · MMLU ↗
73.4% · Claude Instant · MMLU · EM · % · Epoch AI · MMLU ↗
73.2% · Claude Instant · MMLU · EM · % · Epoch AI · MMLU ↗
TriviaQA ↗78.9%
Claude Instant · TriviaQA · EM · % ↗
2 more reported results
78.9% · Claude Instant · TriviaQA · EM · % · Epoch AI · TriviaQA ↗
78.7% · Claude Instant · TriviaQA · EM · % · Epoch AI · TriviaQA ↗
DTBench ↗45.78%
Claude Instant · DTBench · Accuracy · % ↗
Epoch Capabilities Index (ECI) ↗120.21 ECI
Claude Instant · General ECI · 26 October 2026 fit ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
There are not enough shared scored settings to identify close comparisons reliably.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- DTBench — Conceptual Reasoning Index ↗
- Holistic Evaluation of Language Models (HELM) ↗
- AI benchmarks & capabilities ↗
- Model Card and Evaluations for Claude Models ↗
- Releasing Claude Instant 1.2 ↗
Names used in source tables
Claude Instant