David Deutsch Interviews QRAG
Explore deep ideas and your worldview *
* only if you want to.
Choose the number of quoted selections:
Each choice shows an approximate cost per query, then an approximate time and a quality score when those have been measured. Cost assumes 17,000 input tokens. OpenAI reasoning models and Claude Opus 5.5 assume 17,000 output tokens; the other models assume 2,000. Amounts round to the nearest penny, or to the nearest tenth of a penny when the estimate is under one cent. Time is one short question, rounded to the nearest second. A question that quotes many selections usually takes longer and can cost more. Quality is GPQA Diamond divided by GPT-6 Astra’s 96% and multiplied by 100, so 100 is that published score. Input tokens and output tokens are priced in USD per million, rounded to two significant digits or the nearest tenth of a penny. Qwen 3.7 Max: input tokens $1.5 per million, output tokens $4.4 per million. GPT-6 Astra: input tokens $10 per million, output tokens $50 per million. GPT-6 Sol: input tokens $2 per million, output tokens $10 per million. GPT-6 Luna: input tokens $0.1 per million, output tokens $0.5 per million. Claude Opus 5.5: input tokens $4 per million, output tokens $20 per million. DeepSeek V4 Flash: input tokens $0.055 per million, output tokens $0.11 per million. Qwen 3.7 Plus: input tokens $0.32 per million, output tokens $1.3 per million. Kimi K2.6: input tokens $0.95 per million, output tokens $4 per million. GLM 5.2: input tokens $0.65 per million, output tokens $2.0 per million. Nemotron 3 Super: input tokens $0.08 per million, output tokens $0.45 per million.
Select date range:
Answers synthesized from this collection. Expand sources to read the direct quotes.
Or try a question
Sample conversation for design review. Open live chat →