AI Developer Tools — Coding Assistants & Agents
Claude Code leads the comparable cohort at 83.8 (Terminal-Bench 2.1, running Fable). The widest reach — OpenAI SDK at 296M/mo — carries no benchmark. 66 of 100 governance-ready · 18 of 100 benchmarked — the rest em-dash honestly.
100 TOOLSBENCHMARKED 18GOVERNANCE-READY 66/100FASTEST MOVER Letta Code ▲101.2%UPDATED 09 OCT 21:40 UTC
#TOOLSCOREGOVPRICE
1CClaude Code83.8●Free
2CCursor79.3◐$20
3Mmini-swe-agent76.2●Free
4GGemini CLI65.8●Free
LABELED — real, not rankable
·CCodex82.2●Free
·DDROID77.3◐$20
·JJunie CLI71.0●Free
·WWarp61.2●$20
·LLetta Code59.1○$20
·AAbacus AI Desktop58.4◐$10
·GGrok CLI57.3●Free
·GGoose54.3●Free
·OOpenHands51.9●Free
·OOpencode51.7●Free
·GGitHub Copilot43.4◐$10
·SSWE-agent43.4●Free
·AAider28.3●Free
·AAmazon Q Developer CLI24.5●Free
NO PUBLIC BENCHMARK — em-dash is honest
·GGitHub Copilot Chat—◐$10
·GGitHub Copilot in the CLI—●Free
·KKiro IDE—○$20
·OOpenAI Codex CLI—●Free
·DDevin—○$20
·CCline—●Free
·ZZed—●Free
·DDeep Agents Code—●Free
·CClaude Code Agent—●Free
·GGemini Code Assist—◐$19
·MMistral Vibe—●Free
·CCodeRabbit AI Code Review Platform—●$30
·SSonarQube for IDE—●Free
·Kkiota—●Free
·MMiMo Code—●Free
·QQwen Code—●Free
·AAmazon Kiro—○$20
·AAmp—◐$20
·Aanthropic-cli—●Free
·AAuggie—●Free
·Cclaude-code-action—●Free
·Cclaude-plugins-official—●Free
·CCortex Code CLI—●Free
·CCursor Agent—◐$20
·DDevin Desktop/CLI—◐$20
·DDevon—●Free
·GGrok Build—●Free
·KKimi Code—●Free
·PPhpStorm AI Assistant—○$5.42
·PPlandex—●Free
·PPOOL—●Free
·TTabnine CLI—◐$59
·TTRAE IDE—○$3
·VVS Code Server—●Free
·OOpenAI SDK—●Free
·MMCP Python SDK—●Free
·CCursor Cloud Agents—◐$20
·@@modelcontextprotocol/sdk—●Free
·AAnthropic MCP SDK—●Free
·MMCP TypeScript SDK—●Free
·LLangGraph Deep Agents—●Free
·Nn8n—◐$24
·MManus—◐$39
·OOpenAI Agents SDK—●Free
·Oopenai-agents—●Free
·Oopenai-agents-python—●Free
·GGoogle Agent Development Kit—●Free
·CClaude Code SDK—●Free
·LLiveKit Agents Framework—●Free
·SSGLang—●Free
·CClaude Agent SDK—●Free
·LLlamaIndex Agents—◐—
·Llanggraph-sdk—◐$39
·CCrewAI—◐$25
·LLangChain—◐$39
·@@assistant-ui/core—●Free
·NNeMo—○$93.75
·QQwen Agent—●Free
·UUI-TARS-desktop—●Free
·OOpen Interpreter—●Free
·Rrd-agent—●Free
·AAgent Toolkit for AWS—●Free
·AAgents—●Free
·AAmazon Bedrock AgentCore—○—
·AAmazon Bedrock AgentCore harness—○—
·AAutoGen Studio—●Free
·BBedrock Agents—○—
·BBedrock Managed Agents—○—
·BBrowserBase agents—○$20
·Cclaude-agent-sdk-python—◐—
·Ccommerce-agents—●Free
·DDeepSeek Harness—●Free
·DDeepSeek Harness 0.1.1-rc.2—●Free
·GGitLab Duo Agent Platform—◐$1
·HHarvey Agents—●—
·LLangSmith—○$39
·MMicrosoft Agent Framework—●Free
·OOpenAI Agents API—○—
·Oopenai-agents-js—●Free
·Oopenai-cua-sample-app—●Free
·SSemantic Kernel—●Free
·TTrae Agent—○$3
GOV verdict — ● ready (open-local or SOC 2+SSO+0-ret) · ◐ posture stated · ○ /trust pending
C
Claude Code
cli · Anthropic
● Free · BYO-key
7 tiers · Free $0 → Enterprise $20
CAPABILITY — the tool’s own harness
83.8on Terminal-Bench 2.1 running Fable · leads the comparable cohort↗
secondary — 37.7 LiveSWEBench (agentic) · liveswebench · labeled, not comparable↗ as of 2026-08-28
MOMENTUM & SCALE
GOVERNANCE — as the vendor’s /trust page states it
OPEN SOURCERUNS 100% LOCALSOC 2ISO 27001HIPAA
SSO · zero-retention · no-train: not stated — never assumed
PLAN LADDER — sourced allowances↗
Free$0No Claude Code access; chat on web, iOS, Android, and desktop only
Pro$20Claude Code included; usage shared with chat across rolling 5-hour session windows and a weekly cap
Max 5x$1005× the usage allowance of Pro; same 5-hour session + weekly cap structure
Max 20x$20020× the usage allowance of Pro; same 5-hour session + weekly cap structure
Team Standard$251.25× Pro usage per seat; 2-seat minimum, up to 150 seats; Claude Code included; $20/seat billed annually
Team Premium$1256.25× Pro usage per seat; Claude Code + Claude Cowork included; $100/seat billed annually
Enterprise$20
AI Tools — every cell sourced or an honest em-dash, never fabricated149 tools · benchmarks + registries + /trust pages + filings