Claude Haiku 4.5 vs GPT-5.6 Luna
GPT-5.6 Luna is the better pick: it passed more tasks on the AIGROW small-business benchmark and has the lower list price. Claude Haiku 4.5 passed 7 of 20 tasks and costs $1 in and $5 out per million tokens; GPT-5.6 Luna passed 16 and costs $0.20 in and $1.20 out.
AIGROW credit covers Claude Opus 5.5, Claude Fable 5.2, Claude Fable 5.1, GPT-6 Astra, Claude Sonnet 5.5, GPT-6.1 Sol, GPT-6.1 Terra, Grok 4.7 and Gemini 4 Argon. Claude Sonnet 5.5 below list price
$25 a month brings $110 of API credit and cancels any time; $25 once buys $100. For long documents, GPT-5.6 Luna reads up to 1.05M tokens at once and Claude Haiku 4.5 up to 200K. US dollars, standard rate, shortest context band. Cost per pass from the AIGROW small-business benchmark. Newer comparison: Claude Haiku 5.5 vs GPT-6 Luna
- Claude Haiku 4.5
- $6.62
- Tasks passed
- 7 of 20
- GPT-5.6 Luna
- $0.677
- Tasks passed
- 16 of 20
AIGROW small-business benchmark, failed runs included.
Anthropic's previous small model costs five times Luna on input, so it has to earn the difference. GPT-5.6 Luna costs $0.677 per 1,000 passing answers on the AIGROW small-business benchmark, Claude Haiku 4.5 $6.62. 1,000 support replies cost $5.00 on Claude Haiku 4.5 and $1.08 on GPT-5.6 Luna, so Claude Haiku 4.5 costs 4.6 times as much.
What each one costs
The API price per million tokens and the cheapest app plan with each model, from the makers’ own pages. Each name opens that model’s full price page.
- Claude Haiku 4.5
- On the Anthropic API, Claude Haiku 4.5 costs $1 per million input tokens and $5 per million output tokens.
- GPT-5.6 Luna
- On the OpenAI API, GPT-5.6 Luna costs $0.20 per million input tokens and $1.20 per million output tokens. In ChatGPT, the cheapest plan with it is Plus at $20/mo. Plan prices read 22 October 2026.
Side by side
Each maker’s published rate for its standard tier. Cached input is what a repeated prompt prefix costs once the provider has stored it.
| Fact | Claude Haiku 4.5 | GPT-5.6 Luna |
|---|---|---|
| Maker | Anthropic | OpenAI |
| Tier | Small and fast | Small and fast |
| Input, per M tokens | $1.00 | $0.20 |
| Output, per M tokens | $5.00 | $1.20 |
| Cached input | $0.10 | $0.02 |
| Context window | 200K | 1.05M |
| Released | 15 Oct 2025 | 9 Jul 2026 |
| Weights | Closed weights | Closed weights |
What a job costs on each
1,000 runs of each job at the two list prices. The gap moves from job to job because the models price input, output and cached context differently.
| Job | Claude Haiku 4.5 | GPT-5.6 Luna | Cheaper |
|---|---|---|---|
| Support reply3,000 in, 400 out. | $5.00 | $1.08 | GPT-5.6 Luna, 78% less |
| Document summary25,000 in, 1,000 out. | $30.00 | $6.20 | GPT-5.6 Luna, 79% less |
| Agent task80,000 in (70,000 cached), 3,000 out. | $32.00 | $7.00 | GPT-5.6 Luna, 78% less |
The token counts are ours, chosen to look like real work; the prices are the makers’. A model with no published cache price bills the agent task’s repeated context at its full input rate.
The small print
What each base rate leaves out, from the makers’ own pages: surcharges, discounts, promotions and cache pricing.
- Claude Haiku 4.5
- Legacy model, still available; Anthropic lists its retirement as not sooner than 15 October 2026, and the current Haiku is Claude Haiku 5.5. Cache write $1.25 (5 min) or $2 (1 h). Batch API 50% off. Anthropic’s pricing page
- GPT-5.6 Luna
- Nano tier of the 5.6 family. Price cut 80% on 2026-07-30. Requests above 272K input tokens bill 2x input and 1.5x output ($0.40 / $1.80). Cache write $0.25. Batch and Flex 50% off. Data residency adds 10%. OpenAI’s pricing page
Other comparisons
Every head to head on the list that involves Claude Haiku 4.5 or GPT-5.6 Luna.
Compare Claude Haiku 4.5 and GPT-5.6 Luna with 52 other models
The price list sets every model’s input and output price side by side, with what a support reply, a document summary and an agent task cost on each.
Prices checked 12 November 2026