Open-source agent beats Hermes on GAIA with the same local Qwen-3.6-35B setup
SucceededMind · x · 2026-07-25
Orchestration, not bigger weights, appears to be the lever that matters here.
The post cites a benchmark run where Atomic Agent and Hermes Agent were tested on 53 GAIA Level 1 tasks with the same local setup: a 4-bit Qwen-3.6-35B model on an M4 Max.
- Atomic Agent: 69.8% solved in 3h 12m
- Hermes Agent: 58.5% solved in 5h 10m
That means Atomic solved 6 more tasks and finished almost two hours faster.
The claimed reason is a stricter agent loop that avoids wasting compute: it reuses the KV cache with a byte-stable prompt, batches tool calls as JSON, and compresses results instead of stuffing raw logs into context.
More from coding & agent
- He tells AI agents to use the Obsidian CLI instead of grep, mv, and sed — dSebastien · 2026-07-25
- Hands-on workshop on building AI agents pairs with a post on real user changes — hugobowne · 2026-07-25
- Coding benchmarks show agentic models are mostly tackling feature work, bug fixes and optimization — zainhas · 2026-07-25
- Codex Built a Windows Hyper-V VM, Migrated 130 Torrents, and Set Up NordVPN Isolation — Fringolicious · 2026-07-25
- Why a stronger model may work better as designer, with weaker models doing the execution — dotey · 2026-07-25
- 15 Claude Code integrations show how to wire it into coding, testing, docs, and deploys — Aiden_Tech_Ai · 2026-07-25