Home › Compare › GPT-4.1 vs Claude Haiku 4.5
GPT-4.1 vs Claude Haiku 4.5
A straight cost comparison of GPT-4.1 (OpenAI) and Claude Haiku 4.5 (Anthropic), per 1,000,000 tokens, verified October 2026.
Rate-by-rate comparison
| Metric | GPT-4.1 | Claude Haiku 4.5 | Winner |
|---|---|---|---|
| Input / 1M | $2 | $1 | Claude Haiku 4.5 |
| Cached input / 1M | $0.50 | $0.10 | Claude Haiku 4.5 |
| Cache write / 1M | — | $1.25 | GPT-4.1 |
| Output / 1M | $8 | $5 | Claude Haiku 4.5 |
| Context window | 1,047,576 | 200,000 | GPT-4.1 |
| Batch API | Yes | Yes | Tie |
| Long-context tier | None | 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-4.1 | $4 | +$1.75 |
| Claude Haiku 4.5 | $2.25 | 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-4.1 | Claude Haiku 4.5 | Winner |
|---|---|---|---|
| Classification 100K reqs · 500 in / 10 out | $108 | $55 | Claude Haiku 4.5 |
| Data extraction 10K reqs · 2K in / 200 out | $56 | $30 | Claude Haiku 4.5 |
| Support chat 1K reqs · 4K in / 600 out | $12.80 | $7 | Claude Haiku 4.5 |
| Code review 2K reqs · 8K in / 1.5K out | $56 | $31 | Claude Haiku 4.5 |
| Summarisation 1K reqs · 20K in / 1K out | $48 | $25 | Claude Haiku 4.5 |
| Agent loop 5K reqs · 12K in / 2K out | $200 | $110 | Claude Haiku 4.5 |
| Translation 500 reqs · 15K in / 15K out | $75 | $45 | Claude Haiku 4.5 |
| Long-document QA 200 reqs · 120K in / 800 out | $49.28 | $24.80 | Claude Haiku 4.5 |
| Cached RAG 1K reqs · 50K in (45K cached) / 1K out | $40.57 | $14.55 | Claude Haiku 4.5 |
| Nightly batch 10K reqs · 10K in / 2K out | $180 | $100 | Claude Haiku 4.5 |
GPT-4.1 wins 0 of 10, Claude Haiku 4.5 wins 10 of 10. The winner is set by the token mix, not by the headline rate.
Where the two bills cross over
Claude Haiku 4.5 is cheaper on both sides of the mix: $1 vs $2 per million input tokens, and $5 vs $8 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-4.1 | Claude Haiku 4.5 | Difference |
|---|---|---|---|
| 100 | $1.28 | $0.70 | Claude Haiku 4.5 cheaper by $0.58 |
| 1,000 | $12.80 | $7 | Claude Haiku 4.5 cheaper by $5.80 |
| 10,000 | $128 | $70 | Claude Haiku 4.5 cheaper by $58 |
| 100,000 | $1,280 | $700 | Claude Haiku 4.5 cheaper by $580 |
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-4.1 | Claude Haiku 4.5 | Winner |
|---|---|---|---|
| 0% | $12.80 | $7 | Claude Haiku 4.5 |
| 50% | $9.80 | $5.20 | Claude Haiku 4.5 |
| 80% | $8 | $4.12 | Claude Haiku 4.5 |
| 95% | $7.11 | $3.58 | Claude Haiku 4.5 |
Both models publish a cached-read rate, so the discount is comparable: GPT-4.1 reads cache at $0.50 per million, Claude Haiku 4.5 at $0.10.
Which one should you pick?
GPT-4.1 wins on none of the ten workloads above, while Claude Haiku 4.5 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 ($2 vs $1); output rate ($8 vs $5); cached-input rate ($0.50 vs $0.10); cache-write rate (— vs $1.25); context window (1,047,576 vs 200,000).
Batch mode is available on GPT-4.1 and Claude Haiku 4.5, which halves the nightly job. Both publish cached-input discounts, so a stable system prompt is worth caching on either.
FAQ
Is GPT-4.1 cheaper than Claude Haiku 4.5?
On a blended 1M-input-plus-250K-output workload, Claude Haiku 4.5 is cheaper by 44% ($2.25 vs $4). The gap changes for cache-heavy prompts and for batch jobs, as the workload table shows.
Which handles longer prompts, GPT-4.1 or Claude Haiku 4.5?
Claude Haiku 4.5 has the larger context window (1,047,576 vs 200,000 tokens). Neither model publishes a long-context surcharge.
Do both models support prompt caching?
GPT-4.1 bills cached reads at $0.50 per 1M; Claude Haiku 4.5 at $0.10.
Related comparisons
- GPT-5.4 mini vs GPT-4.1
- GPT-6.1 Sol vs GPT-4.1
- Claude Sonnet 5 vs Claude Haiku 4.5
- All OpenAI pricing
- All Anthropic pricing