Home › Compare › GPT-5.5 vs Claude Sonnet 4.6
GPT-5.5 vs Claude Sonnet 4.6
A straight cost comparison of GPT-5.5 (OpenAI) and Claude Sonnet 4.6 (Anthropic), per 1,000,000 tokens, verified October 2026.
Rate-by-rate comparison
| Metric | GPT-5.5 | Claude Sonnet 4.6 | Winner |
|---|---|---|---|
| Input / 1M | $5 | $3 | Claude Sonnet 4.6 |
| Cached input / 1M | $0.50 | $0.30 | Claude Sonnet 4.6 |
| Cache write / 1M | — | $3.75 | GPT-5.5 |
| Output / 1M | $30 | $15 | Claude Sonnet 4.6 |
| Context window | 1,050,000 | 1,000,000 | GPT-5.5 |
| Batch API | Yes | Yes | Tie |
| Long-context tier | > 272,000 | None | — |
For per-token metrics, the lower number wins. For context window and batch support, the higher or broader option wins. A dash means the provider does not publish that rate: a missing cache-write fee means writes are not billed per token, while a missing cached-input rate means cache hits bill at the full input rate.
Blended cost: 1M input + 250K output
| Model | Blended cost | Difference |
|---|---|---|
| GPT-5.5 | $12.50 | +$5.75 |
| Claude Sonnet 4.6 | $6.75 | Cheaper |
Ten workloads, two bills
The same two models, priced across ten workloads that differ only in their input:output ratio. Nothing else changes — no retries, no system prompt, no middleware.
| Workload | GPT-5.5 | Claude Sonnet 4.6 | Winner |
|---|---|---|---|
| Classification 100K reqs · 500 in / 10 out | $280 | $165 | Claude Sonnet 4.6 |
| Data extraction 10K reqs · 2K in / 200 out | $160 | $90 | Claude Sonnet 4.6 |
| Support chat 1K reqs · 4K in / 600 out | $38 | $21 | Claude Sonnet 4.6 |
| Code review 2K reqs · 8K in / 1.5K out | $170 | $93 | Claude Sonnet 4.6 |
| Summarisation 1K reqs · 20K in / 1K out | $130 | $75 | Claude Sonnet 4.6 |
| Agent loop 5K reqs · 12K in / 2K out | $600 | $330 | Claude Sonnet 4.6 |
| Translation 500 reqs · 15K in / 15K out | $263 | $135 | Claude Sonnet 4.6 |
| Long-document QA 200 reqs · 120K in / 800 out | $125 | $74.40 | Claude Sonnet 4.6 |
| Cached RAG 1K reqs · 50K in (45K cached) / 1K out | $77.70 | $43.66 | Claude Sonnet 4.6 |
| Nightly batch 10K reqs · 10K in / 2K out | $550 | $300 | Claude Sonnet 4.6 |
GPT-5.5 wins 0 of 10, Claude Sonnet 4.6 wins 10 of 10. The winner is set by the token mix, not by the headline rate.
Where the two bills cross over
Claude Sonnet 4.6 is cheaper on both sides of the mix: $3 vs $5 per million input tokens, and $15 vs $30 per million output. No input:output ratio reverses that, so on price alone the other model never becomes the better choice — it has to win on context, latency or output quality.
Cost at four scales
The same support-chat workload (4K input, 600 output per request) at four monthly volumes.
| Requests / month | GPT-5.5 | Claude Sonnet 4.6 | Difference |
|---|---|---|---|
| 100 | $3.80 | $2.10 | Claude Sonnet 4.6 cheaper by $1.70 |
| 1,000 | $38 | $21 | Claude Sonnet 4.6 cheaper by $17 |
| 10,000 | $380 | $210 | Claude Sonnet 4.6 cheaper by $170 |
| 100,000 | $3,800 | $2,100 | Claude Sonnet 4.6 cheaper by $1,700 |
What caching does to each bill
Caching only helps the part of the prompt that repeats. This is the support-chat workload (1,000 requests, 4,000 input tokens each) with an increasing share of that prompt served from a warm cache, written once on the first request.
| Cached share of the prompt | GPT-5.5 | Claude Sonnet 4.6 | Winner |
|---|---|---|---|
| 0% | $38 | $21 | Claude Sonnet 4.6 |
| 50% | $29.01 | $15.61 | Claude Sonnet 4.6 |
| 80% | $23.61 | $12.37 | Claude Sonnet 4.6 |
| 95% | $20.92 | $10.75 | Claude Sonnet 4.6 |
Both models publish a cached-read rate, so the discount is comparable: GPT-5.5 reads cache at $0.50 per million, Claude Sonnet 4.6 at $0.30.
Which one should you pick?
GPT-5.5 wins on none of the ten workloads above, while Claude Sonnet 4.6 wins on classification, data extraction, support chat, code review, summarisation, agent loop, translation, long-document qa, cached rag, nightly batch.
Beyond price, the two differ on input rate ($5 vs $3); output rate ($30 vs $15); cached-input rate ($0.50 vs $0.30); cache-write rate (— vs $3.75); context window (1,050,000 vs 1,000,000); long-context tier (yes, above 272,000 vs none).
Batch mode is available on GPT-5.5 and Claude Sonnet 4.6, which halves the nightly job. Both publish cached-input discounts, so a stable system prompt is worth caching on either.
FAQ
Is GPT-5.5 cheaper than Claude Sonnet 4.6?
On a blended 1M-input-plus-250K-output workload, Claude Sonnet 4.6 is cheaper by 46% ($6.75 vs $12.50). The gap changes for cache-heavy prompts and for batch jobs, as the workload table shows.
Which handles longer prompts, GPT-5.5 or Claude Sonnet 4.6?
GPT-5.5 has the larger context window (1,050,000 vs 1,000,000 tokens). Watch the long-context threshold on GPT-5.5: rates rise once input passes it.
Do both models support prompt caching?
GPT-5.5 bills cached reads at $0.50 per 1M; Claude Sonnet 4.6 at $0.30.
Related comparisons
- GPT-5.6 Sol vs GPT-5.5
- GPT-6 Astra vs GPT-5.5
- Claude Sonnet 5 vs Claude Sonnet 4.6
- All OpenAI pricing
- All Anthropic pricing