DigitalOcean's new server-side model synthesis tool on its Inference Engine outperformed Fable 5 at half the cost, scoring 65.65% on the DRACO benchmark at $0.83 per task versus Fable 5's 62.21% at $1…
Merge co-founder and CEO Shensi Ding launched Fusion on July 23rd, a product that lets developers send a single API call to multiple AI models and uses a separate judge model to assemble their respons…
Perplexity AI released WANDR (Wide ANd Deep Research), an open benchmark and evaluation harness built around 500 realistic data-collection tasks for knowledge work. WANDR tests whether research agents…
Perplexity AI open-sourced WANDR, a benchmark for evaluating AI agents on complex research tasks, on July 11, 2026. The benchmark, which stands for Wide and Nuanced Deep Research, measures performance…
Harness-native routing, which optimizes model selection based on the full execution state, is emerging as a key technique for building smarter AI agents, according to a technical analysis. The approac…
OpenRouter launched Fusion, a tool that synthesizes outputs from multiple AI models into a single response, achieving scores of 69.0% on the DRACO deep research benchmark—surpassing individual frontie…