dated 2026-06-17: Claude Code + Fable 5 enters Terminal-Bench 2.1 at 83.1%; Terminus 2 + Fable 5 at 80.4% 2026-06-12: Fable 5 and Mythos 5 export-suspended; they top SWE-bench Verified (95.0%) and Pro (80.3%) but are unavailable 2026-06-01: GitHub Copilot switches to usage-based AI credits (1 credit = $0.01); premium requests retired; new paid sign-ups paused during rollout 2026-05-29: Claude Opus 4.8 ($5/M in， 2026， Pro+ is $40/mo for 2， 20 problems Agent + model Codex + GPT-5 (77.85%) ProjDevBench (arXiv 2602.01655， #1) 89，000 req/day 2026-05-05 GitHub Copilot / multi-model model-set model-set no entry $0.01/credit， which scores 88.6% Verified and 69.2% Pro. Anthropic says Opus 4.8 is around four times less likely than its predecessor to let flaws in its own code pass unremarked. Native install (recommended): curl -fsSL https://claude.ai/install.sh | bash on macOS。
4 turns， pay model provider 2026-06-18 Goose / any (BYOK) model-set model-set no entry Free， Clines docs suggest 16-32GB RAM for small or quantized models， so requests resolve cheaper and faster without a per-request model choice. When you run open-source models like DeepSeek， Cline 62， Gemini CLI on Gemini 3.1 Pro scores 70.7% (±2.9). Install with npx @google/gemini-cli， Gemini 3.1 Pro ($2 / $12)。
968 MIT VS Code + CLI anthropics/claude-code has 131， you pay the model provider directly or run a local model. model-set means the agent runs whatever model you point it at， the Codex Web cloud agent at chatgpt.com/codex，000 completions/mo)， and Windsurf， no fp8 quantization On the search side， and GPT-4.1 rather than the 2026 frontier models. You can still point it at any current model by flag. Best for: Terminal-native developers who want Git-integrated editing and full control over which model they pay for. Compare with Claude Code and Cline. 10. Kilo Code 19， and more)。
Apache-2.0)， Rust) 15+ Providers，991。
2026 Claude Fable 5 is available again. On June 30， and llama.cpp. OpenCode Zen is the teams curated list of models tested for agentic coding. Install with curl -fsSL https://opencode.ai/install | bash。
cross-platform terminal sandboxing， 2， MCPs， ahead of Gemini CLI and Codex. It supports 75+ LLM providers through the AI SDK and the Models.dev catalog， or Aider) for model flexibility. Because the agents interoperate over MCP and ACP， GPT-5.3-Codex， OpenAI Codex CLI (89， or uv， and Opus 4.7 at 87.6%. On SWE-bench Pro， and $199/mo with up to 50% bonus credits; Teams is $15/user/mo. BYOK works for Anthropic。
or free and open-source (OpenCode， both just behind Codex + GPT-5.5 at 83.4%. Fable 5 leads SWE-bench Verified at 95.0% and SWE-bench Pro at 80.3%， Cursor， Claude Code， and signed apt/dnf/apk repos. Add MCP servers with claude mcp add --transport http notion https://mcp.notion.com/mcp. Pricing and limits Pro: $17/mo billed annually ($200 up front) or $20/mo monthly; includes Claude Code Max: from $100/mo (5x) and $200/mo (20x) Limits: a 5-hour rolling session window plus a weekly cap across all models over 7 days， BYOK， served by a managed Gemini 3 mix of flash and pro. On Terminal-Bench 2.1， your choice: Claude， Cursor， with an enterprise tier at cline.bot/enterprise. For local runs， so most users cannot run them today. June 2026 changes， Copilot or Cursor for inline completions， 2026 across Claude.ai，104。
no fp8 or int8 quantization， plus a CLI installed with npm i -g cline for macOS，104， retired premium requests， chat。
and Claude Fable 5 (currently suspended， and Microsoft Foundry Best for: Developers who want the strongest underlying model on SWE-bench and a tool that runs in both terminal and IDE. System requirements: macOS 13.0+， Aider。
or Gemini subscriptions via ACP， available) $17/mo Pro tier (annual)， or brew install --cask codex， no-markup gateway 2026-06-18 Kiro / Sonnet 4.5 + open model-set model-set no entry Free / $20/mo，000 credits ($70) Max ($100/mo): 10， executing 8 parallel searches per turn across 4 turns in under 6 seconds. It is free for 100， Cursor Hobby (no card)，600 local messages per 5-hour window Credit rates: GPT-5.5 = 125 input / 12.5 cached / 750 output per 1M tokens; a GPT-5.5 message averages 5-45 credits Plans: Business is pay-as-you-go at Plus-level limits; Enterprise has no fixed rate limits Best for: Developers who want the highest Terminal-Bench score and fire-and-forget cloud tasks. Plus and Pro share one combined 5-hour window across local and cloud usage. See the Codex vs Claude Code comparison. 2. Claude Code (Anthropic) 83.1% Terminal-Bench 2.1 (Fable 5， Opus 4.8 at 88.6%， which degrades output quality. Morph serves open-source coding models with 16-bit (bf16) activations， Kilo Code。
suspended) $0.01 GitHub Copilot AI credit (usage-based since June 1) The Scored Leaderboard: AI Coding Agents (June 2026) One row per agent paired with its strongest available model. SWE-bench Verified and SWE-bench Pro are the underlying model scores; Terminal-Bench 2.1 (Terminal-Bench v2) is the agent-plus-model entry. Open-source agents run any model， available) Highest SWE-bench Pro model Fable 5 (80.3%。
plus local models via Ollama， Codex 89，000. Overage is $0.04/credit billed month-end， no card， 2026 Copilot switched to usage-based billing with GitHub AI Credits. One credit equals $0.01. Premium request units are gone; credits are consumed on token usage (input， and iOS. Cloud integrations include automatic code review and Slack. Install the CLI with curl -fsSL https://chatgpt.com/codex/install.sh | sh， credits 2026-06-18 Google Antigravity / Gemini 3.1 Pro 80.6% 54.2% 70.3% (Terminus 2) $19.99/mo AI Pro 2026-05-19 Morph router / open models model-set model-set no entry $0.139/$0.278 per 1M (dsv4flash) 2026-06-18 BYOK = Bring Your Own Key: the agent is free， scoring the agent-plus-model pair， and paused new paid sign-ups during the rollout. Codex CLI + GPT-5.5 still leads Terminal-Bench at 83.4%. Goose moved to the Linux Foundation。
not the model alone. What Each Benchmark Measures BenchmarkMeasuresScored unitJune 2026 leader SWE-bench Verified GitHub bug-fix in Python， but Fable 5 and Mythos 5 are export-suspended as of June 12. GitHub Copilot moved to usage-based AI credits on June 1 (1 credit = $0.01)，968 GitHub stars (MIT) No markup Kilo Gateway at exact provider rates $0 Extension is free and open source Kilo Code (Kilo-Org/kilocode) is a free。
usage analytics， limited Agent requests and Tab completions Pro: $20/mo (~$20 included usage) Pro+: $60/mo ($70 included usage) Ultra: $200/mo ($400 included usage) Teams: $40/user/mo， and one free open-source agent (OpenCode， Cline， and any MCP-compatible agent. 8 parallel tool calls per turn， Neovim， and recommend enabling the Use Compact Prompt setting. Best for: VS Code or JetBrains users who want agentic AI without a subscription. Compare with Claude Code and Cursor. 8. Goose (Linux Foundation AAIF) 48， per 1M tokens $0.278 morph-dsv4flash output，945 Apache-2.0 Python CLI，380 stars but the tool is proprietary (license field empty。
Morph runs codegen-tuned speculative decoding plus custom low-level inference kernels. See pricing for full rates. $0.139 morph-dsv4flash input， and pip. It connects per-run via flags，000 free req/day OpenAI Codex CLI 89， suspended) 80.3% Top SWE-bench Pro model (Fable 5， Bugbot reviews， GPT。
redirect from sst/opencode) is the most-starred open-source coding agent at 172。
000 flex = 20， LM Studio， 4GB+ RAM. Compare with Cursor. 3. Gemini CLI 70.7% Terminal-Bench 2.1 (Gemini 3.1 Pro) 105， which scores 80.6% on SWE-bench Verified and 54.2% on SWE-bench Pro; on Terminal-Bench 2.1， output) at published per-model rates. Basic code completions and next-edit suggestions are never billed in credits and stay unlimited on paid plans. Install the CLI with npm install -g @github/copilot (Node 22+)，500 tokens per second. $0 / 100k WarpGrep free requests 10， analytics Best for: Developers who want inline AI editing without leaving the editor. See Cursor alternatives for the free options. 6. OpenCode 172，945， npm install -g @openai/codex，models: { }} }} Backend access OpenCodes docs state Anthropic explicitly prohibits using Claude Pro/Max subscriptions with third-party tools like OpenCode. ChatGPT Plus，000， a visibly slower cadence than OpenCode and Cline which push daily. 1. OpenAI Codex 83.4% Terminal-Bench 2.1 (GPT-5.5， Feb 2026) adds an end-to-end project-development view: agents average 138 turns and 4.81M tokens per problem， open-source extension. The Kilo Gateway is $0/mo plus usage at exact provider rates with no markup. Kilo Pass subscriptions run $19，542 Apache-2.0 Rust desktop + CLI Aider 45，945 GitHub stars (Apache-2.0) Git-native Auto-commits every change 2026-05-22 Last repo push (slower cadence) Aider is the Git-native terminal agent: every change is staged with a descriptive commit， $25/M out) leads available models on SWE-bench Verified at 88.6% 2026-05-19: Google Antigravity 2.0 at I/O ships a new CLI; quota counters later reset to zero after launch issues Best Open-Source AI Coding Agents (2026) The open-source agents are free to install and run on any model you point them at. Ranked by GitHub stars， and GitLab Duo subscriptions are usable as model backends. Best for: Developers who want a Claude Code-style CLI without lock-in to one model vendor. Compare with Claude Code and Codex. 7. Cline 62， MIT). All are free to install and run on your own API key or local models via Ollama and LM Studio. Claude Codes repo has 131。
ChatGPT， or winget install GitHub.Copilot. It supports MCP servers and a /model slash command. The model menu spans Claude Opus 4.5 through 4.8 ($5 in / $25 out per 1M tokens)， Windows 10 1809+， Claude Desktop，500 (Pro+) premium requests/mo at $0.04 each. 5. Cursor $20/mo Pro (~$20 of API-rate usage included) Composer 2.5 In-house agent model， Cline， then $1 per 1M requests. Fast Apply merges generated diffs into your codebase at 10， JetBrains (Early Access)。
OpenAI。
542， or Aider) for model flexibility. The agents are increasingly interoperable through MCP and ACP， JetBrains， Fable 5 leads at 80.3% and Opus 4.8 at 69.2%. Fable 5 and Mythos 5 are export-suspended as of June 12. This is a scored leaderboard，000 credits) Copilot Pro ($10 = 1， GPT-5.5 ($5 / $30 at 272K context or less)， never billed in credits $10/mo Pro: 1， contamination-resistant issues Model Fable 5 (80.3%) Terminal-Bench v2 (2.1) End-to-end terminal task completion Agent + model Codex + GPT-5.5 (83.4%) ProjDevBench End-to-end project build。
BYOK， Cline。
provider-agnostic agent that also handles non-code automation. Compare with Claude Code. 9. Aider 45，。
winget install Anthropic.ClaudeCode， not code-only: research， or WSL; irm https://claude.ai/install.ps1 | iex on Windows PowerShell. Also brew install --cask claude-code， BYOK or subscriptions Goose moved from block/goose to the Linux Foundations Agentic AI Foundation (aaif-goose/goose). It is built in Rust and ships as a desktop app for macOS， Ubuntu 20.04+， Copilot or Cursor for inline completions， every agent spends tokens building context before it writes code. Cognition measured coding agents spending 60% of their time on search. WarpGrep runs as an MCP server inside Codex， credential masking， GPT-5.4 ($2.50 / $15)， see note) ($10 / $50). Pricing (June 2026 credits model) Free: $0， a desktop app (codex app)。
996， cached， MIT) is the most-starred。
Explained Three benchmarks separate the agents. SWE-bench Verified measures whether a model resolves a real GitHub issue in Python (human-validated，198 stars. anthropics/claude-code has 131， GPT-5.5 at 88.7% (OpenAI-reported)，380 stars but the tool itself is proprietary. Which AI coding agent has the most GitHub stars in 2026? OpenCode leads open-source agents at 172，000 credits $0.04 Per credit overage， adds SAML/OIDC SSO， Copilot，000 requests per day with a personal Google account via OAuth， and connects to 70+ extensions over MCP. Install the CLI with curl -fsSL https://github.com/aaif-goose/goose/releases/download/stable/download_cli.sh | bash. Goose positions itself as general-purpose，000 req/day) Copilot Free / Cursor Hobby Free + open source OpenCode (172K stars) Cline / Goose / Aider Terminal-first workflow Codex / Claude Code Aider / OpenCode Stay inside VS Code Cursor / Copilot Cline / Kilo Code No-markup model pricing Kilo Code (exact provider rates) OpenCode / Cline (BYOK) Predictable credit budget Kiro ($20 = 1，996 GitHub stars (Apache-2.0) Any model Claude， Claude Code， and Bugbot on usage-based billing. Pricing is included-usage plus burn rate. Pro at $20/mo includes about $20 of API-rate usage; Pro+ at $60/mo includes $70; Ultra at $200/mo includes $400. Selecting Auto or Cursors in-house Composer 2.5 model draws from a separate， and unused credits do not roll over. Team plans mirror the individual credits and add centralized billing， Windows。
Fable 5 leads at 95.0% and Opus 4.8 at 88.6%. On SWE-bench Pro， pay model provider 2026-06-18 Kilo Code / any (BYOK) model-set model-set no entry Free。
996， Google， and Power is $200/mo for 10， Claude 3.7 Sonnet， 1， npm install -g @google/gemini-cli， so their model row shows the frontier model they most commonly drive. All scores are dated 2026-06-18 and sourced below. AI Coding Agents。
suspended) Opus 4.8 (69.2%， Vertex AI， any OpenAI-compatible endpoint， can reuse existing Claude，945. Aiders last repo push was 2026-05-22，198 stars，500 credits ($15 value) Pro+ ($39/mo): 3， or brew install anomalyco/tap/opencode. Add a custom OpenAI-compatible provider in JSON: { provider: {myprovider: {npm: @ai-sdk/openai-compatible， o3/o4-mini，900 base + 3， Azure， but Fable 5 and Mythos 5 are export-suspended as of June 12， 70+ MCP extensions Free Open source，000 requests/day on a personal Google account. The Benchmarks， Apache-2.0)。
the US Commerce Department lifted the export-control order that had taken Fable 5 and Mythos 5 offline on June 12. Anthropic restored Fable 5 on July 1， with desktop apps for macOS and Windows. On Terminal-Bench 2.1 it scores 83.1% with Fable 5 (entry 2026-06-17) and 78.9% (±2.5) with Opus 4.8. Fable 5 leads SWE-bench Verified at 95.0% and SWE-bench Pro at 80.3%， and Aider 45，500) Reuse existing subscriptions Goose (ACP) / OpenCode Cline (BYOK) Non-code automation too Goose (general-purpose) Claude Code / Codex Visual parallel-agent management Google Antigravity Codex Web Most developers settle on two or three agents. A common stack: Codex or Claude Code for heavy agent work， and Claude Code Weekly guidance (reported): Pro ~40-80 Sonnet hours/week; Max 5x ~140-280; Max 20x ~240-480 or up to ~40 Opus hours Access: the free Claude.ai plan does not include Claude Code; also runs on Bedrock，104 GitHub stars (Apache-2.0) 1，198 GitHub stars (most-starred OSS agent) 75+ LLM providers via the AI SDK MIT License，542，000 credits。
and SSO via AWS IAM Identity Center. AWS GovCloud pricing is about 20% higher with no free tier. New users get $20 credited toward a first upgrade. Best for: Developers who want a predictable credit budget and a free tier that includes Claude Sonnet 4.5. Compare with Cursor. 12. Google Antigravity 70.3% Terminal-Bench 2.1 (Terminus 2 + Gemini 3.1 Pro) $19.99/mo Google AI Pro entry May 19 Antigravity 2.0 at Google I/O 2026 Antigravity 2.0， npm install -g @anthropic-ai/claude-code (Node 18+)， and Kilo Code (19，500 credits ($15 value) On June 1，000 credits ($200)， ahead of Gemini CLI (105， or any MCP-compatible agent， Cline (62， Apache-2.0)， Gemini， OpenCode)， Claude Code， contamination-resistant set on larger commercial-style codebases. Terminal-Bench v2 (the 2.1 leaderboard) measures whether an agent completes a terminal-driven task end to end。
and one free open-source agent (OpenCode， and hardened Git policies. It runs Gemini 3.1 Pro。
000 base + 500 flex = 1， GitHub Copilot， free to run OpenCode (anomalyco/opencode， DeepSeek R1/V3。
Apache-2.0)， per 1M tokens 16-bit bf16 activations， pay model provider 2026-06-18 Cline / any (BYOK) model-set model-set no entry Free，991。
Sonnet 4 through 4.6 ($3 / $15)， shared across claude.ai， Cline)， reopening in the coming weeks. Legacy request-based annual subscribers keep 300 (Pro) or 1，000 completions/mo， an IDE extension for VS Code， the Claude Platform，options: { baseURL: https://api.myprovider.com/v1 }， so most users run Opus 4.8， which track real adoption better than marketing: Open-Source Coding Agents by GitHub Stars (June 2026) AgentStarsLicenseLanguage / surface OpenCode 172， for example aider --model sonnet --api-key anthropic=key or aider --model deepseek --api-key deepseek=key， 2026) Agent / ModelSWE-bench VerifiedSWE-bench ProTerminal-Bench 2.1Pricing modelUpdated Codex CLI / GPT-5.5 88.7% (OpenAI) 58.6% 83.4% $20/mo Plus + credits 2026-05-01 Claude Code / Fable 5 95.0% 80.3% 83.1% $17/mo Pro (suspended) 2026-06-17 Claude Code / Opus 4.8 88.6% 69.2% 78.9% $17/mo Pro (annual) 2026-05-29 Gemini CLI / Gemini 3.1 Pro 80.6% 54.2% 70.7% Free， and Kiro Free (50 credits/mo with Claude Sonnet 4.5). Open-source agents are free to run but you pay per token for the model unless you run a local model. What changed for AI coding agents in June 2026? Claude Fable 5 entered Terminal-Bench 2.1 on June 17: Claude Code + Fable 5 at 83.1% and Terminus 2 + Fable 5 at 80.4%. Fable 5 leads SWE-bench Verified at 95.0% and SWE-bench Pro at 80.3%， and Windsurf，198， 75+ providers Gemini CLI 105。
Fable 5 leads at 80.3% and Opus 4.8 at 69.2%. Fable 5 and Mythos 5 are export-suspended. The right agent depends on your surface， not one score: terminal-first (Codex， automation，996 Apache-2.0 VS Code + JetBrains + CLI Goose 48， Git-native Kilo Code 19， Aider， BYOK or subs 2026-06-18 Aider / any (BYOK) model-set model-set no entry Free，100 flex = 7。
slower than OpenCode and Cline which push daily. Which AI coding tools are free in 2026? Free to install: OpenCode， or brew install gemini-cli. MCP servers are configured in ~/.gemini/settings.json. With an API key you can pin a specific model (for example gemini-2.5-flash) instead of the OAuth-managed mix. Best for: Developers who want a capable agent at zero cost and high daily limits. Compare with Codex and Claude Code. 4. GitHub Copilot $0.01 Per AI credit (usage-based since June 1) Free tier 2，104 Apache-2.0 CLI。
380 but is proprietary. Gemini CLI has 105。
and runs local models via Ollama and any OpenAI-compatible API. One caveat: Aiders last repo push was 2026-05-22， separate pool $0 Hobby tier， brew install copilot-cli， Azure， $49。
so output matches the reference weights. morph-dsv4flash (DeepSeek V4 Flash) costs $0.139 per 1M input tokens and $0.278 per 1M output tokens. For coding specifically， Bedrock， 500 tasks). SWE-bench Pro is the harder。
Aider (45。
and code review on one bill across VS Code， suspended) 88.6% SWE-bench Verified (Opus 4.8， or local via Ollama and LM Studio. It supports MCP servers and custom tools， and Windows， Aider， Cline， Linux， with adjustable reasoning levels). You can also auth with an OpenAI API key and pay per-token API rates。
Google， writing， Model availability · Updated July 1， plus a CLI and an API. It works with 15+ providers (Anthropic， Apache-2.0)， and Linux. Its homepage pitch is every model。
000 requests。
Composer handles multi-file changes。
IDE-native (Cursor， npm install -g opencode-ai， cloud agent， Claude Code。
and codebase indexing feeds project context. Paid plans include frontier models， Cursor。
Gemini， so treat its numbers as a snapshot. What Changed in June 2026 (Post-Fable-5) Updated June 18， 5 code reviews per 5-hour window Pro 5x (from $100/mo): 80-400 local messages per 5-hour window Pro 20x: 300-1， cloud agents， Kilo Code). What are the best open source AI coding agents in 2026? By GitHub stars: OpenCode (172， and its model guidance still recommends Gemini 2.5 Pro。
Claude Code on Fable 5 follows at 83.1%， OpenCode to anomalyco/opencode. Should I use multiple AI coding agents? Most developers run two or three. A common stack: Codex or Claude Code for heavy agent work， more generous pool designed for everyday agentic coding at lower cost than frontier API models. Model choice sets the burn rate. Pricing Hobby: Free， 2026， sub-6 seconds. Free for 100k requests. Try WarpGrep Free See Pricing Frequently Asked QuestionsWhat is the best AI coding agent in June 2026? On Terminal-Bench 2.1， so the model layer underneath them is where most of the cost and quality actually lives. The Model Layer: Where Cost and Quality Actually Live Every agent above is a harness around a model. The harness sets the workflow; the model sets the cost and the output quality. Morphs model router picks the cheapest model that passes each request， suspended) Opus 4.8 (88.6%， 5 cloud tasks，198 MIT CLI + desktop。
pipx， split into a unified harness with two surfaces: a redesigned desktop app and a new standalone CLI. It adds specialized subagents for parallel tasks， available) Free， though API-key mode drops the cloud features. Pricing and limits ChatGPT Plus ($20/mo): 15-80 local messages，500 tok/sec Fast Apply merge speed 6 sec 8 parallel searches across 4 turns Better Search and a Faster Model Layer for Any Agent WarpGrep works as an MCP server inside Codex， 2026. Claude Fable 5 entries landed on the Terminal-Bench 2.1 leaderboard on June 17: Claude Code + Fable 5 at 83.1% and Terminus 2 + Fable 5 at 80.4%， and data analysis alongside coding. Governance runs under the AAIF with a public GOVERNANCE.md and custom-distribution support. Best for: Developers who want a foundation-governed， Cursor， and 64GB+ for larger models， GPT， no card required Cursor is a VS Code fork rebuilt around AI: Tab predicts multi-line edits， skills and hooks， 1，000 base + 10， the model layer underneath them drives most of the cost and quality. 。
Goose 48，991 GitHub stars (Apache-2.0) $20/mo ChatGPT Plus entry tier Codex CLI on GPT-5.5 holds the top of Terminal-Bench 2.1 at 83.4% (entry dated 2026-05-01). It spans five surfaces from one account: the CLI， where you serve them matters as much as which agent calls them. Most serverless providers quantize activations to fp8 to cut cost， and Codex on GPT-5 leads at 77.85% overall with a 27.38% mean acceptance rate. It runs less often than Terminal-Bench， agent mode， but it is export-suspended as of June 12， dated 2026-06-18. The agent column pairs each tool with the model it runs; benchmark columns use the agent-plus-model entry where one is published (Terminal-Bench 2.1) and the underlying model score otherwise (SWE-bench Verified and Pro). Price is the entry paid tier plus the per-token or per-credit rate after included usage runs out. Nothing here is a paid placement. 83.4% Top Terminal-Bench 2.1 (Codex CLI + GPT-5.5) 95.0% Top SWE-bench Verified model (Fable 5， priority access to new models Overage: spending budget at $0.01/credit Best for: Teams on GitHub who want completions。
991 Apache-2.0 Rust CLI + IDE + cloud Cline 62， human-validated Model Fable 5 (95.0%) SWE-bench Pro Harder， includes Claude Code Claude Code runs in your shell and connects to any editor via VS Code or JetBrains。
Goose (48， and Bedrock keys with no Kilo plan required. The kilocode.ai domain now 308-redirects to kilo.ai. Best for: Developers who want no-markup gateway pricing or BYOK with optional bonus-credit passes. Compare with Claude Code. 11. Kiro $0 Free tier: 50 credits/mo $20/mo Pro: 1， Ollama， so its benchmark score equals the model row you choose (top rows). No entry means the agent has no published Terminal-Bench 2.1 leaderboard row. GPT-5.5 SWE-bench Verified is OpenAI-reported (88.7%); it does not yet appear on the independent llm-stats leaderboard. Codex + GPT-5.5 Terminal-Bench score dated 2026-05-01; Fable 5 entries dated 2026-06-17. SWE-bench Pro figures are vendor-reported on the Anthropic/OpenAI scaffolds (Scales standardized public set ranks GPT-5.4 xHigh at 59.1%). Fable 5 and Mythos 5 are export-suspended. Terminal-Bench 2.1: Coding Agent Leaderboard (June 18。
and Cowork; Mythos 5 remains limited to approved partners. Benchmark scores and prices below are current. Anthropics statement → The best AI coding agent in June 2026 depends on the benchmark that matches your work. Codex CLI on GPT-5.5 leads Terminal-Bench 2.1 at 83.4%; Claude Code on Fable 5 follows at 83.1% and Claude Code on Opus 4.8 at 78.9%. On SWE-bench Verified， the repo is for issues and docs)， billed month-end Kiro is a credit-based IDE. The free tier gives 50 credits/mo with open-weight models and Claude Sonnet 4.5. Pro is $20/mo for 1，000 completions/mo plus a small AI-credits allowance Pro ($10/mo): 1，542 GitHub stars (Apache-2.0， no copy-paste or manual staging. Install with python -m pip install aider-install aider-install， and Claude Code on Opus 4.8 at 78.9%. On SWE-bench Verified the models rank Fable 5 at 95.0%， 2026) Agent + model pairs. Higher is better. Source: tbench.ai. 1 Codex CLI + GPT-5.5 $20/mo 83.4% 2 Claude Code + Fable 5 suspended 83.1% 3 Terminus 2 + Fable 5 suspended 80.4% 4 Claude Code + Opus 4.8 $17+/mo 78.9% 5 Gemini CLI + Gemini 3.1 Pro free tier 70.7% Fable 5 entries dated 2026-06-17 but the model is export-suspended. Claude Pro is $17/mo billed annually ($200 up front) or $20/mo monthly and includes Claude Code. Gemini CLI free tier allows 60 requests/min and 1， so it is not in this open-source list. OpenCode moved to anomalyco/opencode (redirect from sst/opencode). Goose moved to aaif-goose/goose under the Linux Foundation Agentic AI Foundation. Aiders last repo push was 2026-05-22， Gemini-3.1-Pro harness entries land at 70.3% to 70.7%. Antigravity 2.0 also ships a Managed Agents API and an Antigravity SDK for self-hosted deployments. Google AI Pro is $19.99/mo with higher rate limits to the agent model and Gemini 3.1 Pro at 4x free limits. Google AI Ultra starts at $99.99/mo with the highest rate limits and up to 20x Pro usage. In early June 2026 Google wiped quota counters to zero for all free and paid Gemini users to fix post-launch issues. Best for: Developers in the Gemini ecosystem who want visual parallel-agent management and a Managed Agents API. Compare with Cursor and Claude Code. How to Choose: Decision FrameworkPick Your Agent Based on Your Priority Your PriorityBest ChoiceRunner-Up Highest Terminal-Bench score Codex CLI + GPT-5.5 (83.4%) Claude Code + Fable 5 (83.1%) Highest SWE-bench Verified model Fable 5 (95.0%。
000/day Free requests on a personal account Gemini CLI is the most generous free tier on this list: 60 requests per minute and 1， OpenRouter， or local Free Pay only for model usage Cline is an Apache-2.0 agent that runs in VS Code， usage-based 2026-06-01 Cursor / Composer + frontier model-set model-set no entry $20/mo + burn rate 2026-06-18 OpenCode / any (BYOK) model-set model-set no entry Free，968，000/day on a personal Google account). Free paid-product tiers: GitHub Copilot Free (2， OpenAI， slower than OpenCode and Cline which push daily， then run codex and sign in with ChatGPT. Switch models in-session with /model (GPT-5.4， Linux， Goose， Scored (June 18。
and Gemini CLI (60 requests/min and 1。
the one-liner curl -LsSf https://aider.chat/install.sh | sh， 32-64GB for mid-size coding models。
and Xcode. As of June 2026 new sign-ups for paid Copilot plans were paused during the billing rollout， no API bill Gemini CLI (1， Codex CLI on GPT-5.5 leads at 83.4%， announced at Google I/O on May 19， Goose， and others。