Together AI boosted agent task success from 67% to 96% with homegrown agent evals
nutlope · x · 2026-11-17
Together AI's CEO details how they made their platform easier for coding agents, raising task success from 67% to 96% by building their own agent evals. They benchmarked baseline vs MCP vs Skills across 12+ tasks (two runs each, incomplete runs count as failures): Skills hit 97.2% pass rate, cut incomplete rate from 9.7% to 2.8%, and reduced avg end-to-end time from 11.6 to 9.3 minutes at $0.433/run. Full methodology included for building your own agent evals.
More from coding & agent
- EMBRACE: Embedded Browsers in Cursor, Copilot, Codex, VS Code, Claude Desktop All Hacked — cyb3rops · 2026-11-17
- OpenCode Browse 0.1.0 Rewritten in Rust: 25x Less Memory, Playwright Dropped — gaganghotra_ · 2026-11-17
- Together AI's agent eval workflow lifted task completion from 67% to 96% — togethercompute · 2026-11-17
- Codex and Claude MCP Wired into DaVinci Resolve Hailed as a Standout Workflow — _AustinCalvert_ · 2026-11-17
- Beaker tunes your agent harness per model before you pick one — VictorSanh · 2026-11-17
- VapiCon 2026 takeaways: phone calls still dominate voice AI, sims and evals go mainstream — chadwallacehart · 2026-11-17