Balancing the Anvil

These are the models on the anvil, the frontier. Each runs with a simple riff-*, so you can focus on the model and the work at hand.

Model Coverage

Default Harness
Alternate Harness
Unavailable
Which model each CLI runs, and the riff command that launches it.
Model Claude Junie Codex Antigravity
Anthropic / Claude
Fable 5 default harnessriff-fable alternate harness unavailableunavailable
Opus 5 default harnessriff-opus alternate harness unavailableunavailable
Sonnet 5 default harnessriff-sonnet alternate harness unavailableunavailable
Haiku 4.5 default harnessriff-haiku unavailableunavailableunavailable
OpenAI / GPT
GPT-5.6 Sol unavailable alternate harness default harnessriff-sol unavailable
GPT-5.6 Terra unavailable alternate harnessriff-junie default harnessriff-terra unavailable
GPT-5.6 Luna unavailable alternate harness default harnessriff-luna unavailable
Google / Gemini
Gemini 3.1 Pro unavailable alternate harness unavailable default harnessriff-gemini-pro
Gemini 3.6 Flash unavailable alternate harness unavailable default harnessriff-gemini-flash

Effort per Model

The reasoning effort each riff targets, beside the vendor's default.
Model Vendor Default Anvil Target Change
Anthropic / Claude
Fable 5highhighunchanged
Opus 5highxhighraised
Sonnet 5highxhighraised
Haiku 4.5unavailable
OpenAI / GPT
GPT-5.6 Sollowlowunchanged
GPT-5.6 Terramediummediumunchanged
GPT-5.6 Lunamediummediumunchanged
Google / Gemini
Gemini 3.1 Prohighhighunchanged
Gemini 3.6 Flashmediummediumunchanged

Cost

Every tier keeps the JetBrains IDEs; each step up the ladder adds more capability to the anvil. Two ladders climb the same seven tiers, one focused on Anthropic and the other on OpenAI. They are alternatives, not stages. Pick the ladder focused on the model you reach for most, since the money buys depth in one vendor rather than savings.

Which vendor the ladder focuses on
Cost by tier and vendor, with Anthropic in focus. Each vendor row merges the tiers a plan covers into one span; a colored left edge marks the tier where a plan upgrades. Each cell leads with the monthly and shows the yearly beneath; the totals row gives an effective monthly and the yearly.
VendorStarterApprenticeJourneymanProfessionalEliteUltimateMythic
AnthropicFreeClaude Pro≈$17/mo$200/yrClaude Max 5x$100/mo$1,200/yrClaude Max 20x$200/mo$2,400/yr
OpenAIFreeChatGPT Plus$20/mo$240/yrChatGPT Pro 5x$100/mo$1,200/yrChatGPT Pro 20x$200/mo$2,400/yr
GoogleFree
JetBrainsAll Products Pack≈$15/mo$179/yr
Total≈$15/mo$179/yr≈$32/mo$379/yr≈$115/mo$1,379/yr≈$135/mo$1,619/yr≈$235/mo$2,819/yr≈$315/mo$3,779/yr≈$415/mo$4,979/yr
The same seven tiers with OpenAI in focus instead. Each vendor row merges the tiers a plan covers into one span; a colored left edge marks the tier where a plan upgrades. Each cell leads with the monthly and shows the yearly beneath; the totals row gives an effective monthly and the yearly.
VendorStarterApprenticeJourneymanProfessionalEliteUltimateMythic
AnthropicFreeClaude Pro≈$17/mo$200/yrClaude Max 5x$100/mo$1,200/yrClaude Max 20x$200/mo$2,400/yr
OpenAIFreeChatGPT Plus$20/mo$240/yrChatGPT Pro 5x$100/mo$1,200/yrChatGPT Pro 20x$200/mo$2,400/yr
GoogleFree
JetBrainsAll Products Pack≈$15/mo$179/yr
Total≈$15/mo$179/yr≈$35/mo$419/yr≈$115/mo$1,379/yr≈$132/mo$1,579/yr≈$232/mo$2,779/yr≈$315/mo$3,779/yr≈$415/mo$4,979/yr

*Anthropic Claude Pro bills at $20 per month or $200 per year. This anvil example assumes the annual rate, which saves $40.

*JetBrains All Products Pack is $299 the first year, $239 the second, then $179 onwards. This anvil example qualifies for the $179 rate.

Add-ons (optional at any tier, not counted in the totals above)
Optional add-ons. These attach to any tier and are not included in any total above. The monthly and yearly columns give each add-on's own cost.
VendorPlanMonthlyYearly
GoogleAI Pro+$20+$240
AI Ultra+$100+$1,200
JetBrainsAI Ultimate+$25+$300

*Google AI Pro costs $19.99 per month and AI Ultra $99.99, rounded here to $20 and $100 for simplicity.

*JetBrains AI Ultimate is billed yearly at $300.

The riff bill of rights

What is a riff-*? A short shell command, named after a musical riff, a phrase you play again and again. A riff bakes in your model of choice and any additional arguments you want. You can have as many as you like; think of them as shortcuts you reach for often.

An example riff-* setup in bash
# the shared engine every riff launcher calls: run the CLI (blocking), then # restore your working directory when the CLI exits, via a trap _riff() { local bin="${1}" origin="${PWD}"; shift # never turn an agent loose in your home dir (ssh keys, dotfiles, creds); sandbox it # defaults to ~/sandbox; override with AI_SHELL_SANDBOX if [[ "${PWD}" == "${HOME}" ]]; then cd "${AI_SHELL_SANDBOX:-${HOME}/sandbox}" || return; fi # ~~~~~ make it yours here: customizations, tab title, color, whatever ~~~~~ trap 'trap - RETURN; cd "${origin}"' RETURN # back to where you started when the CLI exits "${bin}" "${@}" # keep last so a Ctrl-C still runs the cleanup } # each riff launcher is just a thin declaration: a client, a model, and any flags you want riff-opus() { _riff claude --model opus --effort xhigh; } riff-sol() { _riff codex --model gpt-5.6-sol --config "model_reasoning_effort='low'"; } riff-gemini-pro() { _riff agy --model "Gemini 3.1 Pro (High)"; }