Claude Opus 4.8 fast mode in GitHub Copilot
Anthropic - Powerful- preview
In GitHub Copilot, Claude Opus 4.8 fast mode costs $10 per 1M input tokens and $50 per 1M output tokens, billed as AI credits at $0.01 each. A typical daily-coding request works out to about 12.7 credits (about $0.13) with auto model selection on. Rates verified against GitHub's rate sheet on 2026-07-14.
Token rates
- Input / 1M
- $10.00
- Cached input / 1M
- $1.00
- Output / 1M
- $50.00
- Cost per credit
- $0.01
Cache writes bill at $12.50/1M tokens (Anthropic models).
Credits per request
Defined workload baskets, not measured tasks, with auto model selection on.
| Workload | Credits (auto on) | Credits (auto off) | Cost |
|---|---|---|---|
| Short chat turn | 6.5 | 7.2 | $0.06 |
| Daily coding request | 12.7 | 14.1 | $0.13 |
| Agent session step | 41.9 | 46.6 | $0.42 |
At about 12.7 credits per daily-coding request, Claude Opus 4.8 fast mode ranks #22 of 22 current models by cost, about 44.1x the cheapest (GPT-5.4 nano).
Requests per plan allowance
Daily-coding requests the included monthly credits buy, before any overage.
| Plan | Credits / month | Requests |
|---|---|---|
| Copilot Pro | 1,500 | 118 |
| Copilot Pro+ | 7,000 | 551 |
| Copilot Max | 20,000 | 1,576 |
| Copilot Business | 1,900 | 149 |
| Copilot Enterprise | 3,900 | 307 |