Fable 5.1
Anthropic
Validation index85.1Cost / task$2.260Time / task6.8
Agentic Test & Validation
Visual comparisons for engineering workflows, using the final benchmark’s validation index, cost per task, and task duration.
Highlights
Overall indexLeading models at a glance
Overall index
Performance landscape
Every model plotted against performance, so the tradeoff is visible at a glance — not just the ranking.
How to read this chart+
Each point is one model, colored by provider — click a provider in the legend below the chart to isolate it. The dotted line connects the Pareto frontier: models where no other model scores at least as well on both axes and strictly better on one. The shaded corner marks the most attractive combination of cost and performance.
Overall index
Top 12Performance across models
Category-normalized validation index. Higher is better.
Overall index
13 modelsAll model results
Every model in the final run, ranked by validation index, with cost and task duration alongside.
GPT-6 Astra
OpenAI
Validation index85.1Cost / task$2.260Time / task6.7
Opus 5
Anthropic
Validation index83.4Cost / task$1.130Time / task6.7
GPT-5.6 Sol
OpenAI
Validation index81.6Cost / task$0.452Time / task6.3
Sonnet 5
Anthropic
Validation index79.0Cost / task$0.452Time / task6.1
GPT-5.6 Terra
OpenAI
Validation index78.1Cost / task$0.514Time / task6.0
Kimi K2.7 Code
MoonshotAI
Validation index72.9Cost / task$0.150Time / task6.2