Field noteNov 16, 20265 min read1 source

Models and pricing for GitHub Copilot

In one minute

GitHub’s Copilot pricing page lists Gemini 4 Hydrogen as generally available, with rates of $1.50 per million input tokens, $0.075 per million cached input tokens, and $7.50 per million output tokens. For teams choosing models inside Copilot, the entry makes both its status and metered costs explicit, while reminding them that paid usage depends on plan allowances and budget controls.

What happened

GitHub’s “Models and pricing for GitHub Copilot” documentation describes token-based pricing for models in Copilot. It says an interaction can consume input tokens sent to a model, output tokens generated by it, and cached tokens that the model reuses or stores. Each token is priced according to the model, and the resulting cost is converted into GitHub AI credits; one credit equals $0.01 USD. The page lists Gemini 4 Hydrogen in the Google section as GA, in the Versatile category and Default tier. Its per-million-token prices are $1.50 for input, $0.075 for cached input and $7.50 for output. No separate long-context threshold appears for Hydrogen in that table.

The documentation says that the way usage is tracked and billed depends on the Copilot plan. Copilot Free, Pro, Pro+ and Max include individual AI-credit allowances that vary by plan. Business and Enterprise include per-user allowances pooled at the billing-entity level. If a plan’s included allowance is exceeded, additional use is billed in AI credits at the per-token rates in the tables. The page does not specify a Hydrogen-specific allowance or say which individual or organizational plans can select it; its stated rates describe the metered cost after included allowances.

GitHub also documents a way to prevent spending beyond a plan’s included credits: create or edit a Bundled AI credits budget, set it to $0 USD and enable “Stop usage when budget limit is reached.” The page notes a mobile-purchase complication: people who bought Pro, Pro+ or Max through an iOS or Android in-app purchase cannot buy additional credits through the app store. If a payment method is on file with GitHub, additional paid usage is billed to that method based on budget settings. GitHub says users can move billing management to GitHub.com by cancelling the mobile subscription and resubscribing there.

The pricing page places Hydrogen among other Google models, including Gemini 3.8 Flash, Gemini 4 Argon and Gemini 4 Carbon. It also states that code completions and next-edit suggestions remain unlimited on paid Copilot plans and are not billed in AI credits. That distinction matters when estimating costs: the listed token rates describe model interactions, not every Copilot feature. The documentation does not give a Hydrogen-specific example of a typical interaction or resulting bill.

What the archive adds

New Runtime’s earlier coverage said Gemini 4 Carbon was generally available in GitHub Copilot for Pro+, Max, Business and Enterprise users, with usage-based billing (Gemini 4 Carbon in GitHub Copilot, November 10, 2026). This is a new model listing and pricing fact, not a restatement of that Carbon coverage. An earlier note also covered Hydrogen arriving in Cursor (Gemini 4 Hydrogen is now available in Cursor!, November 12, 2026); this page adds a distinct Copilot billing context.

Trend signal

This adds another example of model choice inside a coding assistant being paired with per-token costs rather than only a plan label. It weakly supports the idea that agent economics is moving toward the cost of completed work, but GitHub’s page gives token rates, not cost per verified task. The distinction is central to Agent economics moves to completed work: builders still need workload measurements to connect a model’s unit price to engineering outcomes.

Why now

GitHub’s documentation presents Hydrogen’s status and rates alongside other models and explains how overage billing works. That makes the model’s cost legible at the point where teams compare options, although the page does not state when the entry was added or why it changed.

Why it matters

A model’s headline capability is only one part of choosing it for coding work. The rate difference between input, cached input and output means that prompt structure, repeated context and generated response length can all affect spend. For teams using Copilot beyond plan allowances, those costs flow through AI credits and can be pooled at an organization’s billing entity, so a model-level choice may become a team budget issue.

The page also draws an important boundary: code completions and next-edit suggestions are not billed in AI credits, while model interactions are metered. A team should not treat its Copilot usage as one uniform cost category. It should identify which workflows consume credits, confirm allowance and billing ownership, and set a stop budget if unplanned overages are unacceptable. The published prices enable estimates, but not a conclusion about which model is cheapest for a particular repository or task.

Limits and uncertainty

The pricing documentation lists Hydrogen as GA, but it does not state when it became selectable in Copilot or enumerate the plans and clients that can use it. It gives per-million-token rates, not measured costs for typical coding tasks, and does not report quality or latency comparisons against other models. Included credit allowances vary by plan, but the Hydrogen entry does not identify a model-specific quota. Actual spend therefore depends on token volumes, cached-context behavior, plan terms and any budget settings; the listed rate alone does not establish a team’s likely bill.

Reader takeaway

Gemini 4 Hydrogen is listed as GA in GitHub Copilot at $1.50 per million input tokens and $7.50 per million output tokens, with a lower cached-input rate. Before using it beyond included credits, check the applicable plan allowance and set a budget stop if overages must be prevented.

Evidence / sources

  1. 1docs.github.com/en/copilot/reference/copilot-billing/models-and-pricing
All Field Notes

Next reads

Read next

Open the graph
  1. 01topicGithubExplore the github topic hub.
  2. 02topicGeminiExplore the gemini topic hub.
  3. 03topicGoogleExplore the google topic hub.
  4. 04archiveField NotesOpen the latest editorial analysis.
  5. 05sourcesSourcesSee the original sources this archive cites.
Who read this page?Machine requests, hidden until opened

Loading the privacy-safe route aggregate...

Open the JSON contract