Claude Haiku 5.5
Long-context pricing tiers
| Tier | Input | Cached input | Output |
|---|---|---|---|
| Long context | $0.50 | $0.05 | $2.50 |
| Short context | $0.10 | $0.01 | $0.50 |
Batch API: input $0.05 and output $0.25 per 1M tokens.
Claude Haiku 5.5 vs GPT-6 Luna API Pricing
Compare verified published API prices for Claude Haiku 5.5 and GPT-6 Luna. This page compares price, not output quality.
Verified Oct 21, 2026, Claude Haiku 5.5 lists input at $0.10 and output at $0.50 per 1M tokens, while GPT-6 Luna lists $0.10 input and $0.50 output. They have the same published list price for both token types; their published long-context pricing tier differs, and this comparison does not assess output quality.
| Model | Input $/1M | Cached input $/1M | Output $/1M | Source | Last verified |
|---|---|---|---|---|---|
| Claude Haiku 5.5 | $0.10 | $0.01 | $0.50 | Official source | Oct 9, 2026 |
| GPT-6 Luna | $0.10 | $0.01 | $0.50 | Official source | Oct 21, 2026 |
Prices from official provider pages, verified on the date shown.
Published rates
Side-by-side API pricing
| Field | Claude Haiku 5.5 | GPT-6 Luna |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Input / 1M | $0.10 | $0.10 |
| Cached input / 1M | $0.01 | $0.01 |
| Output / 1M | $0.50 | $0.50 |
| Batch input / 1M | $0.05 | $0.05 |
| Batch output / 1M | $0.25 | $0.25 |
| Context window | 1M | 1050K |
| Release date | Oct 7, 2026 | Sep 22, 2026 |
| Official source | Anthropic pricing ↗ | OpenAI pricing ↗ |
Standard token prices
Which is cheaper depends on your input:output ratio
Costs below allocate 1M total tokens across each input:output ratio. They use standard prices only, without cache or Batch discounts.
| Input:output ratio | Claude Haiku 5.5 | GPT-6 Luna | Cheaper published price |
|---|---|---|---|
| 10:1 (RAG / long context) | $0.136 | $0.136 | Same price |
| 3:1 (chat) | $0.20 | $0.20 | Same price |
| 1:1 | $0.30 | $0.30 | Same price |
| 1:3 (long generation) | $0.40 | $0.40 | Same price |
Computed examples
Monthly cost for real workloads
| Workload | Assumed volume | Claude Haiku 5.5 | GPT-6 Luna | Difference |
|---|---|---|---|---|
| Side project | 10,000 requests/month | $2.30 | $2.30 | Same cost |
| Customer support chatbot | 10,000 conversations/month | $17.57 | $17.57 | Same cost |
| Agent task | 1,000 tasks/month | $2.80 | $2.80 | Same cost |
Assumptions
- Side project: 10,000 requests per month, with 800 input and 300 output tokens per request.
- Customer support chatbot: 10,000 conversations, with 6 replies per conversation, 600 system tokens, 90 user-message tokens, 1,000 retrieved-context tokens, 180 output tokens per reply, and 50% history carry.
- Agent task: 1,000 tasks, with 8 calls per task and 1,500 input plus 400 output tokens per call.
- All three rows use standard non-cached, non-batch prices and exclude retries, tools, storage, and taxes.
Provider conditions
Pricing rules that can change the answer
GPT-6 Luna
Long-context pricing tiers
| Tier | Input | Cached input | Output |
|---|---|---|---|
| Long context | $0.20 | $0.02 | $0.75 |
| Short context | $0.10 | $0.01 | $0.50 |
Batch API: input $0.05 and output $0.25 per 1M tokens.
Verified changes
Price history
On Oct 7, 2026, Anthropic added Claude Haiku 5.5 with published input $0.10 and output $0.50 per 1M tokens.
View this update in the price monitorOn Sep 22, 2026, OpenAI replaced GPT-5.6 Luna with GPT-6 Luna: input $0.20 → $0.10 (−50%) and output $1.20 → $0.50 (−58.3%) per 1M tokens.
View this update in the price monitor
Operational costs
Cheaper per token is not always cheaper per task
Retries, failed calls, fallback requests, and longer outputs can raise the cost of completing a task even when the published token price is lower. Model the effect with the LLM Retry & Fallback Cost Calculator, read the cost-per-successful-task guide, or estimate a migration with the Model Switch Cost Calculator.
Primary evidence and next steps
Related
Common questions
Frequently asked questions
Is Claude Haiku 5.5 cheaper than GPT-6 Luna?
Input prices are the same. Output prices are the same. This compares published list prices, not output quality.
How much do Claude Haiku 5.5 and GPT-6 Luna cost for 10,000 requests a month?
At 800 input and 300 output tokens per request, Claude Haiku 5.5 costs about $2.30 and GPT-6 Luna costs about $2.30 at standard list prices.
Which is cheaper for input-heavy workloads like RAG?
At a 10:1 input-to-output token ratio, Claude Haiku 5.5 costs $0.136 and GPT-6 Luna costs $0.136 per 1M total tokens. Same price.
Where do these prices come from?
The prices come from Anthropic and OpenAI official pricing sources, verified on Oct 9, 2026 and Oct 21, 2026, respectively. See the AI API pricing dataset.