Top 15 of 16 scored · 21 benchmarks · method v1
Agents: 1. GPT-6 Astra, 2. Claude Fable 5.1, 3. Claude Opus 5
Leader
100
GPT-6 Astra · OpenAI
Best open weights
n/a
No scored model
Best value
63
Gemini 3.7 Flash · Google DeepMind$3.75 per 1M output tokens