A Vercel benchmark for AI coding agents on Next.js code generation and migration tasks, reporting success rate, average execution time, and an AGENTS.md documentation-assisted split.
BenchLM mirrors the public Next.js Evals agent leaderboard from Vercel. The source evaluates 18 model and coding-agent harness rows on 24 Next.js code generation and migration tasks, reporting base success rate, average duration, and a separate success rate when agents have bundled Next.js documentation through AGENTS.md.
Next.js Evals is display only on BenchLM. It is useful for framework-specific web development work, but each row combines a base model with a coding-agent harness such as Cursor, Claude Code, Codex, Gemini CLI, or OpenCode, so BenchLM keeps it out of weighted model-only rankings.
BenchLM mirrors the published success rate view for Next.js Evals. Composer 2.5 leads the public snapshot at 92% , followed by Claude Opus 4.8 (88%) and GPT-5.5 Pro (83%). BenchLM does not use these results to rank models overall.
Composer 2.5
Cursor
Cursor Composer 2.5 / Cursor
Claude Opus 4.8
Anthropic
Claude Opus 4.8 / Claude Code
GPT-5.5 Pro
OpenAI
GPT 5.5 Pro / Codex
The published Next.js Evals snapshot is tightly clustered at the top: Composer 2.5 sits at 92%, while the third row is only 9.0 points behind. The broader top-10 spread is 17.0 points, so the benchmark still separates strong models even when the leaders cluster.
18 models have been evaluated on Next.js Evals. The benchmark falls in the Coding category. This category carries a 20% weight in BenchLM.ai's overall scoring system. Next.js Evals is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
Year
2026
Tasks
24 Next.js code generation and migration tasks
Format
Agent task completion with withheld Vitest assertions
Difficulty
Framework-specific web application engineering
Next.js Evals focuses on framework-specific web engineering tasks such as Pages Router to App Router migration, server actions, cache directives, proxy middleware, async cookies and headers, and other current Next.js patterns. BenchLM mirrors the public leaderboard as display-only because rows combine model choice with an agent harness.
Version
Next.js Evals 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public tasks with withheld Vitest assertions
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
A Vercel benchmark for AI coding agents on Next.js code generation and migration tasks, reporting success rate, average execution time, and an AGENTS.md documentation-assisted split.
Composer 2.5 currently leads the published Next.js Evals snapshot with 92% success rate. BenchLM shows this benchmark for display only and does not use it in overall rankings.
18 AI models are included in BenchLM's mirrored Next.js Evals snapshot, based on the public leaderboard captured on June 1, 2026.
For engineers, researchers, and the plain curious — a weekly brief on new models, ranking shifts, and pricing changes.
Free. No spam. Unsubscribe anytime.