8 Aug 2026: Claude Sonnet 5 permanent pricing. $2/$10 forever.
Original Plan vs Permanent Pricing
June 30: Sonnet 5 launch. $2/$10 introductory pricing through Aug 31. Sept 1 scheduled increase to $3/$15. Aug 10: Anthropic updates official blog. $2/$10 becomes permanent standard pricing. Sept 1 price increase cancelled.
GitHub Copilot Model Deprecation
Same Sept 1 date: Copilot removes 6 models. Removed: Sonnet 4.6, Opus 4.5, Opus 4.6, Sonnet 4.5, Gemini 3.1 Pro, Raptor Mini. Replacements: Sonnet 4.6 to Sonnet 5, Opus models to Opus 4.7/4.8. Key: migration is price decrease, not increase. Enterprise teams get 33% cost reduction.
Budget Recalculation Scenarios
Scenario A: Planned Sonnet 5 migration. Budgeted $3/$15 Sept 1 increase. Now reimburse 33% savings. Reallocate to other AI initiatives. Scenario B: Avoided price increase. Sonnet 4.6 cost already lower than $3/$15. Must migrate to Sonnet 5 by Sept 1 (Copilot mandatory). Scenario C: Opus 4.8 investment. Sonnet 5 performance now 80% of Opus 4.8 at 40% cost. Implement tiered model: Sonnet 5 for routine, Opus 4.8 for high-difficulty only.
Cost Comparison: Sonnet 5 vs Opus 4.8
10B monthly token scenario. Sonnet 5: 6B input ($2) + 4B output ($10) = $520k/month. Opus 4.8: 6B input ($5) + 4B output ($25) = $1.3M/month. Monthly savings: $780k. Cost difference matters for enterprise scale. Pilot testing required to confirm Sonnet 5 achieves target performance. Effort adjustment: Sonnet 5 at 3x effort approaches Opus 4.8 quality on some tasks.
Model Selection for Q4 Agent Rollout
Choose Sonnet 5 for: routine task automation, real-time bot responses, cost-sensitive departments, high-volume workflows. Choose Opus 4.8 for: complex reasoning, regulated industries, mission-critical decisions, high-stakes analysis. Hybrid approach: Sonnet 5 for initial 80% of workload. Promote to Opus 4.8 if benchmark fails. This two-tier strategy minimizes budget overruns while maintaining quality.
Before Sept 1: Actions Required
1. Audit Q4 budgets. Find $3/$15 Sonnet 5 entries. Calculate 33% savings impact. 2. Identify Copilot Sonnet 4.6 users. Plan migration to Sonnet 5. 3. Enable Sonnet 5 in Copilot Enterprise model policies (manual step). 4. Run benchmark: Sonnet 5 vs Opus 4.8 on real internal tasks. Measure quality, cost, speed. 5. Draft two-tier model policy for Sept 1 rollout. 6. Update spend controls in Claude API and Enterprise dashboards. Getting these six steps right before Sept 1 protects both budget and performance. If your team needs structured guidance on model migration, cost optimization, and Agent readiness, for a rapid assessment and rollout plan.
