Perplexity's WANDR Eval: GPT-6 Astra Tops at 0.682, $11.98 Per Task
cameronstow · x · 2026-09-04
Perplexity published its evaluation of GPT-6 Astra on its in-house benchmark WANDR: the model scored 0.682 — the highest of any model they tested — at $11.98 per task.
Comparisons: GPT-6-Astra scored 13.5% higher than Fable 5.1 at 6.1% lower cost, and 27.0% higher than Opus 5 at just 3.3% higher cost. A systematic third-party capability-vs-price evaluation useful for model selection.
Related event: GPT-6 Astra lands on Perplexity, tops WANDR benchmark(3 posts)→
More from Models
- 3 months testing whether Anthropic's internal research model can design scorpion antivenom — apoorvasriniva · 2026-10-21
- Space Bunny Alpha revealed as MiniMax M3.1 Flash: 1M context, budget-friendly coding — ZabihullahAtal · 2026-10-21
- 'Opus 5.5 Medium' Spotted in Rork's Model Picker, Hinting at New Anthropic Model — rudrank · 2026-10-21
- RSI-Bench Clears First Task Set Through Reward-Hacking Detection Pipeline — SRSchmidgall · 2026-10-21
- Dev complains Anthropic's bio safeguards trigger just from reading a bio paper — robleclerc · 2026-10-21
- Redditor claims uncensored local DeepSeek build answers virus-writing prompt with zero hesitation — Only_Light_9817 · 2026-10-21