TL;DR — Sonnet 5 beats GPT-5.4 on both sides ($2/$10 vs $2.50/$15), and as of Sept 2026 that rate is permanent rather than promotional. Every number below is from our weekly-verified pricing tracker (verified 2026-09-13).
Pricing side by side (per 1M tokens, USD)
| Model | Provider | Input | Output |
|---|---|---|---|
| Claude Sonnet 5 | Anthropic | $2 | $10 |
| GPT-5.4 | OpenAI | $2.5 | $15 |
Verified 2026-09-13 against official rate cards — the tracker keeps the change log. Note: Claude Sonnet 5 — introductory $2/$10 made permanent; the planned Sept 1, 2026 rise to $3/$15 was cancelled.
What a real month costs
At a standard workload of 10M input + 2M output tokens/month:
- Claude Sonnet 5: $40.00
- GPT-5.4: $55.00
On this mix, Claude Sonnet 5 is the cheaper column (1.4× gap). Your ratio will differ — run your own mix through the LLM cost calculator.
Choose Claude Sonnet 5 if
you want strictly cheaper with frontier-adjacent quality — Sonnet 5 wins the price column outright, with no expiry attached.
Choose GPT-5.4 if
you need OpenAI-side features or ecosystem fit — on raw rates GPT-5.4 no longer has a pricing argument here.
Routing decision card
- Sonnet 5 wins the price column outright, and the rate is permanent — but still check the tracker before trusting any comparison you read
- This is the pair where a dated price table matters most; most stale comparisons on the web still describe the $2/$10 as a promo that expired
- Router play: Sonnet as the default mid-tier, GPT-5.4 only where OpenAI-side features matter
- The router answer: most teams shouldn't hard-pick one — ClawRouters routes each request to whichever side fits it, so the comparison becomes a per-call decision instead of a bet.
Part of the LLM model comparison index. Prices move fast in 2026 — the tracker logs every change with dates.