Cursor's CursorBench 4.0: Gemini 4 Carbon tops at 60.4%, DeepSeek under $1 per task
mattyp · x · 2026-11-18
Cursor released CursorBench 4.0, evaluating agent models on ambiguous multi-file tasks from real Cursor sessions. Gemini 4 Carbon High leads at 60.4% ($9.22/task); Fable 5.2 Max scores 60.0% but costs $32.79/task across 218 steps. Value picks: Opus 5.5 High hits 56.0% at $3.97/task, while DeepSeek V4.1 Flash 1017 runs $0.60–0.98/task at 44.9–47.7%. Top scores are tightly clustered while per-task costs vary 60x.
More from Models
- Asking Gemini 10,701 questions: Google Search rewrites ~45% of brand picks — dejanseo · 2026-11-18
- Salesforce's SOV-Transcribe-1 ASR model trends on Hugging Face — salesforce-sovereign-intelligence · 2026-11-18
- Test: Google's AI Mode default Gemini 4 Hydrogen knows almost nothing about brands without search — dejanseo · 2026-11-18
- OpenAI explains the 6.2 influencer/community vibe gap: demo videos used pre-release models — i_dg23 · 2026-11-18
- Ruff author Charlie Marsh: GPT-6.2 was an enormous step-up for my work — charliermarsh · 2026-11-18
- Claude Code ignored an explicit rule and invented timestamps 4 hours ahead — shymecw · 2026-11-18