Models/claude-sonnet-5-5

Claude Sonnet 5.5

First on Android Bench 2.0 long-horizon at 40.0% (claude-code). First on sudo L7 at 50.6% Adaptive/Max. Third on WebDev at 1770. 5th on Supabase without skills at 77%, 86% with skills.

Anthropic · Proprietary · Context 1M · $2 / $10 per million tokens · Fetched 2026-10-10

Android Bench

40.0%

#1 / 11 · 24–56

sudo L7

50.6%

#1 / 16

LMArena WebDev

1770

#3 / 153 · ±11

Benchmark scores

Boards older than 90 days are not counted as recommendations

BenchmarkSettingScoreRankCostAs of

Android Benchclaude-code40.024–56#1$1574.92026-10-26

sudo L7Adaptive/Max50.6#1—2026-10-28

LMArena WebDevxhigh1770±11#3—2026-10-27

LMArena Image-to-WebDevxhigh1737±17#2—2026-10-19

Terminal-Bench 4.0max · Claude Code61.8±2.9#2$7.3k2026-10-06

The Slop Index—68#12—2026-10-28

Supabase Evalsmedium · Claude Code77.0#5—2026-10-30

Convex Evals4 runs85.49#6$1.452026-10-28

FrontierCodemax · claude-code41.0#4$19.942026-10-27

SWE-bench Verified—This board has stopped updating——2026-02-26

Aider Polyglot—This board has stopped updating——2025-11-20

Use it or not

Why

Leads Android long-horizon and sudo L7; for frontend, about 40 fewer WebDev points than Opus at half the price.

When not

Android long-horizon averages 70.7 hours and about $1575.
A Terminal-Bench run costs $7.3k, more than the leader's $4.7k.
Trails Opus on the Supabase no-skills baseline (77% vs 90%).

Appears in these scenarios

Collected from scenario boards

Build a Web frontend#1sudo L7 50.6%Build an Android app#1sudo L7 50.6%Write CLI, scripts, and DevOps#2Terminal-Bench 61.8%Write docs & copy#12Slop Index 68

Appears in these builds

Build a Web frontend with this setClaude Opus 5.5 + frontend-design + shadcn/ui + Phosphor Icons + Motion + Geist + Figma MCP server→Build an Android app with this setClaude Sonnet 5.5 + Jetpack Compose + Android AGENTS.md→

Peers

ModelAndroid BenchPrice $/M

Claude Sonnet 5.5Current40.0%$2 / $10 Claude Opus 5.537.3%$4 / $20 GPT-6.1 Sol29.3%$2 / $10