GitHub Copilot pricing, plans and credits
As of 2026-09-14, GitHub Copilot is free for individuals with limits, $10 a month for Pro (1,500 AI credits, $15 of usage), $39 for Pro+ (7,000 credits), $100 for Max (20,000 credits), $19 per seat for Business (1,900 credits) and $39 per seat for Enterprise (3,900 credits). One credit is $0.01, and credits are consumed by input, output and cached tokens at the model rates GitHub publishes, which match the providers' list prices row for row where a registry rate exists.
Plans and monthly credit allotments
| Plan | Price | AI credits a month | Usage at $0.01 a credit | Note | Source |
|---|---|---|---|---|---|
| Copilot Free | Free | Allowance not itemised | Unpublished | The plans page says Free includes an allowance of GitHub AI Credits without stating the number. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Student | Free | Allowance not itemised | Unpublished | Credit allowance not itemised on the plans page. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Pro | $10/month | 1,500 (1,000 base + 500 flex) | $15 | Access to selected models and agents. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Pro+ | $39/month | 7,000 (3,900 base + 3,100 flex) | $70 | Access to premium models and agents. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Max | $100/month | 20,000 (10,000 base + 10,000 flex) | $200 | Priority access to premium models and agents. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Business | $19 per seat/month | 1,900 | $19 | Plans page states 1,900 credits total per user per month; base and flex split not shown. | GitHub Copilot plans Read 2026-09-14 |
| Copilot Enterprise | $39 per seat/month | 3,900 | $39 | Plans page states 3,900 credits total per user per month; base and flex split not shown. | GitHub Copilot plans Read 2026-09-14 |
The usage column is arithmetic (credits × $0.01), not a GitHub figure. A $10 Pro plan carries $15 of model usage at published rates; the difference is the flex allotment GitHub added on top of the base credits.
How AI credits work
One credit is one cent, consumed by tokens
Interactions consume input tokens, output tokens and cached tokens; each token's cost depends on the model used. Rates per model are in the table below and equal the providers' list prices where a registry rate exists.
Base plus flex, reset monthly
Base credits are included with the plan each month and stay constant; the flex allotment is an additional monthly amount on top of base credits that GitHub says is designed to adapt as the economics of AI evolve. Unused credits are forfeited and the allowance resets to the full monthly amount at 00:00:00 UTC on the first day of each calendar month.
When the allotment runs out
When credits are exhausted, upgrade the plan, set a budget and pay for additional usage at 1 AI credit = $0.01, or wait for the next month. Source: GitHub Copilot billing for individuals, read 2026-09-14.
Per-model rates beside the providers' list prices
Copilot's rate per million tokens, then the provider's list rate from the BenchLM pricing registry (updated September 22, 2026). Parity is computed at build time, never typed: 26 of 27 comparable default-tier rows match the registry on both input and output; differences on Gemini 3.6 Flash; no registry rate yet for MAI-Code-1.1-Flash. Long-context and fast-mode rows are Copilot tiers with no single registry counterpart.
| Model | Tier | Copilot in / cached / out | List in / cached / out | Input | Output |
|---|---|---|---|---|---|
| GPT-5 miniOpenAI | default | $0.25 / $0.025 / $2 | $0.25 / Unpublished / $2 | same | same |
| GPT-5.3-CodexOpenAI | default | $1.75 / $0.175 / $14 | $1.75 / Unpublished / $14 | same | same |
| GPT-5.4OpenAI | default | $2.5 / $0.25 / $15 | $2.5 / $0.25 / $15 | same | same |
| GPT-5.4OpenAI | long-context | $5 / $0.5 / $22.5 | $2.5 / $0.25 / $15 | not comparable | not comparable |
| GPT-5.4 miniOpenAI | default | $0.75 / $0.075 / $4.5 | $0.75 / $0.075 / $4.5 | same | same |
| GPT-5.4 nanoOpenAI | default | $0.2 / $0.02 / $1.25 | $0.2 / $0.02 / $1.25 | same | same |
| GPT-5.5OpenAI | default | $5 / $0.5 / $30 | $5 / $0.5 / $30 | same | same |
| GPT-5.5OpenAI | long-context | $10 / $1 / $45 | $5 / $0.5 / $30 | not comparable | not comparable |
| GPT-5.6 LunaOpenAI | default | $0.2 / $0.02 / $1.2 | $0.2 / $0.02 / $1.2 | same | same |
| GPT-5.6 LunaOpenAI | long-context | $0.4 / $0.04 / $1.8 | $0.2 / $0.02 / $1.2 | not comparable | not comparable |
| GPT-5.6 SolOpenAI | default | $4 / $0.4 / $20 | $4 / $0.4 / $20 | same | same |
| GPT-5.6 SolOpenAI | long-context | $8 / $0.8 / $30 | $4 / $0.4 / $20 | not comparable | not comparable |
| GPT-5.6 TerraOpenAI | default | $2 / $0.2 / $12 | $2 / $0.2 / $12 | same | same |
| GPT-5.6 TerraOpenAI | long-context | $4 / $0.4 / $18 | $2 / $0.2 / $12 | not comparable | not comparable |
| GPT-6 AstraOpenAI | default | $10 / $1 / $50 | $10 / $1 / $50 | same | same |
| GPT-6 AstraOpenAI | long-context | $20 / $2 / $75 | $10 / $1 / $50 | not comparable | not comparable |
| Claude Haiku 4.5Anthropic | default | $1 / $0.1 / $5 | $1 / $0.1 / $5 | same | same |
| Claude Sonnet 4Anthropic | default | $3 / $0.3 / $15 | $3 / Unpublished / $15 | same | same |
| Claude Sonnet 4.6Anthropic | default | $3 / $0.3 / $15 | $3 / Unpublished / $15 | same | same |
| Claude Opus 4.7Anthropic | default | $5 / $0.5 / $25 | $5 / Unpublished / $25 | same | same |
| Claude Opus 4.8Anthropic | default | $5 / $0.5 / $25 | $5 / Unpublished / $25 | same | same |
| Claude Opus 5Anthropic | default | $5 / $0.5 / $25 | $5 / $0.5 / $25 | same | same |
| Claude Sonnet 5Anthropic | default | $2 / $0.2 / $10 | $2 / $0.2 / $10 | same | same |
| Claude Opus 4.8 (fast mode)Anthropic · Fast mode is a separate Copilot rate; the registry row is the standard Opus 4.8 rate | fast-mode | $10 / $1 / $50 | $5 / Unpublished / $25 | not comparable | not comparable |
| Claude Fable 5Anthropic | default | $10 / $1 / $50 | $10 / $1 / $50 | same | same |
| Claude Fable 5.1Anthropic | default | $10 / $0.25 / $50 | $10 / $0.25 / $50 | same | same |
| Gemini 3.5 FlashGoogle | default | $1.5 / $0.15 / $9 | $1.5 / $0.15 / $9 | same | same |
| Gemini 3.6 FlashGoogle · Promotional pricing through December 31, 2026 | default | $0.75 / $0.075 / $3.75 | $1.5 / $0.15 / $7.5 | lower | lower |
| Gemini 3.7 FlashGoogle · Promotional pricing through December 31, 2026 | default | $0.75 / $0.075 / $3.75 | $0.75 / $0.075 / $3.75 | same | same |
| Gemini 3.8 FlashGoogle · Promotional pricing through December 31, 2026 | default | $0.75 / $0.075 / $3.75 | $0.75 / $0.075 / $3.75 | same | same |
| MAI-Code-1.1-FlashMicrosoft | default | $0.2 / $0.02 / $1.2 | Unpublished / Unpublished / Unpublished | no registry rate | no registry rate |
| Grok 4.5xAI | default | $2 / $0.5 / $6 | $2 / $0.3 / $6 | same | same |
| Grok 4.5xAI | long-context | $4 / $1 / $12 | $2 / $0.3 / $6 | not comparable | not comparable |
| Grok 4.6xAI | default | $2 / $0.5 / $6 | $2 / $0.5 / $6 | same | same |
| Grok 4.6xAI | long-context | $4 / $1 / $12 | $2 / $0.5 / $6 | not comparable | not comparable |
| Kimi K2.7 CodeMoonshot AI | default | $0.95 / $0.19 / $4 | $0.95 / Unpublished / $4 | same | same |
| Kimi K3Moonshot AI | default | $3 / $0.3 / $15 | $3 / $0.3 / $15 | same | same |
Credits calculator
Credits your workload would use
Pick a model and a workload shape. Tokens are priced at Copilot's published per-model rates and converted at one credit per cent; each plan row shows whether its monthly allotment covers the month.
- Credits per month
- 185
- $1.85 of usage at one credit per cent
- Split
- $0.35 in · $1.50 out
- 1,800,000 input and 750,000 output tokens on GPT-5 mini
- Same tokens on the provider API
- $1.85
- Copilot bills the provider's list rate per token, so the direct-API figure is the same arithmetic; the plan price is what buys the allotment.
| Plan | Allotment | Covers the month? | Additional usage | Plan + additional |
|---|---|---|---|---|
| Copilot Pro $10.00/month | 1,500 credits | Yes | $0.00 | $10.00 |
| Copilot Pro+ $39.00/month | 7,000 credits | Yes | $0.00 | $39.00 |
| Copilot Max $100/month | 20,000 credits | Yes | $0.00 | $100 |
| Copilot Business $19.00/month | 1,900 credits | Yes | $0.00 | $19.00 |
| Copilot Enterprise $39.00/month | 3,900 credits | Yes | $0.00 | $39.00 |
Keep these assumptions with the decision. The estimate above is today’s list rates against the workload you typed; both move. Save it as a Radar cost scenario and the next review starts from the same numbers, with the source of every rate attached.
Estimate at GitHub's published rates. It excludes fast-mode multipliers, long-context tiers, Spark and cloud-agent SKUs, taxes and any organisation budget cap. Copilot Free and Student have no published credit number and are not listed. Confirm the model identifier in the model ID directory.
One worked session, synthetic: a code-generation month of 300 requests at 6,000 input and 2,500 output tokens with a quarter of input cached uses about 1,179 credits on GPT-5.6 Terra, inside Pro's 1,500. The same month on GPT-6 Astra is roughly five times that and exceeds Pro, which is the arithmetic behind the "credits run out in days" reports: the flat fee used to absorb reasoning-model sessions and now meters them at list price.
What changed, with dates
- 2026-04-27
GitHub announced that all Copilot plans move to usage-based billing: premium requests are replaced by a monthly allotment of GitHub AI Credits consumed by input, output and cached tokens at the published API rate for each model. The announcement listed Pro at $10 of credits, Pro+ at $39, Business at $19 per user and Enterprise at $39 per user. GitHub Blog: Copilot is moving to usage-based billing
- 2026-06-01
Usage-based billing took effect. Annual subscribers keep request-based pricing until their plan expires, then move to a monthly plan or Copilot Free; the legacy premium-request multipliers (code review at 13 from June 1) apply to them until then. GitHub Docs: premium requests (legacy)
- 2026-09-14
The plans page now shows base and flex allotments: Pro 1,000 base plus 500 flex (1,500 credits, $15 of usage on a $10 plan), Pro+ 3,900 plus 3,100, Max 10,000 plus 10,000. GitHub describes the flex allotment as an additional monthly amount designed to adapt as the economics of AI evolve. GitHub Copilot plans
Sources
- 01GitHub Copilot plansRead 2026-09-14
- 02GitHub Copilot models and pricingRead 2026-09-14
- 03GitHub Copilot billing for individualsRead 2026-09-14
- 04GitHub Blog: Copilot is moving to usage-based billing (2026-04-27)Read 2026-09-14
Questions
Is GitHub Copilot free?
There is a Copilot Free plan with limited access to a selection of features, 2,000 completions a month and an allowance of AI credits that GitHub's plans page does not quantify. Verified students get Copilot's features free on the Student plan. Paid plans start at $10 a month for Pro. Read 2026-09-14.
How much does GitHub Copilot cost?
Pro is $10 a month with 1,500 AI credits, Pro+ $39 with 7,000, Max $100 with 20,000; Business is $19 per granted seat with 1,900 credits per user and Enterprise $39 per seat with 3,900. One credit is $0.01, so Pro includes $15 of model usage, Pro+ $70, Max $200, Business $19 and Enterprise $39.
What is a GitHub AI credit worth?
One AI credit equals $0.01. Credits are consumed by the input, output and cached tokens each interaction uses, at the rate GitHub publishes for the model. Unused credits are forfeited and the allowance resets to the full monthly amount at 00:00 UTC on the first day of each month. Additional usage past the allowance is billed at the same rate against a budget you set.
What changed on June 1, 2026?
Premium requests were replaced by AI credits consumed by tokens at published API rates. GitHub announced it on April 27, 2026 and it took effect on June 1. Annual subscribers keep request-based pricing until their plan expires. The plans page has since added a flex allotment on top of base credits, so Pro carries 1,500 credits on a $10 plan.
Is Copilot cheaper than Claude Code or Cursor?
It depends on the tokens your sessions use, because Copilot bills the same per-token rates the providers publish. A $10 Pro plan buys $15 of model usage at those rates; a reasoning-heavy session can use that in days. Our Cursor and Claude Code comparison pages cover features; use the calculator on this page for the credits your workload would consume.