GPT-6 Sol hits #6 on Agent Arena with +7.7% net improvement at 56% less cost per task
arena · x · 2026-09-26
Agent Arena's new leaderboard puts GPT-6 Sol (Max) at #6 with +7.7% net improvement across 4,000+ real-world agentic sessions, up from #8 for its predecessor. Versus GPT 5.6 Sol (xHigh), it gains 1.5 points of net improvement at half the per-token price, and jumps to #4 on Confirmed Success (+11.4% vs +2.9%).
On the cost-performance Pareto frontier, GPT-6 Sol sits at $0.75 median cost/task — 56% cheaper than Claude Fable 5 (High) at $1.72 — while trailing it by only 0.6 percentage points. The arena measures tool orchestration on long-horizon tasks with web, filesystem, and terminal tools across 2M+ sessions.
More from coding & agent
- SAG2 builds RAG index 7000x faster with no embeddings or GPU, beats HippoRAG 2 — udmrzn · 2026-11-11
- One sentence makes Claude Fable 5.2 refactor boldly: say the codebase is its own — brandon_galang · 2026-11-11
- Visa tested 22 models on 15 shopping tasks: ~43% failure, agents bought wrong items anyway — Thick_Mulberry_2235 · 2026-11-11
- MCP or CLI? Yes — Sentry ships a unified experience bridging both — zeeg · 2026-11-11
- Stanford Launches CS329Z: Engineering AI Agents, Covering Memory, Evals and Proactive Agents — stanfordnlp · 2026-11-11
- Solo dev took down his multi-agent site for 8 hours by draining the shared model budget — wolfdogKC · 2026-11-11