Claude Haiku 5.5 vs GPT-5.6 Luna

As of October 21, 2026, GPT-5.6 Luna costs 100% more per input token than Claude Haiku 5.5 ($0.20 vs $0.10 per 1M) and has a 922K-token context window versus 1M.

Prices from LiteLLM's price fileChecked October 21, 2026

Claude Haiku 5.5 vs GPT-5.6 Luna side by side

Headline prices in US dollars per 1M tokens: each model's own lab where it sells it, otherwise its cheapest provider. Source: LiteLLM model price file, as of October 21, 2026.
	Claude Haiku 5.5	GPT-5.6 Luna
Input, per 1M tokens	$0.10	$0.20
Output, per 1M tokens	$0.50	$1.20
Cache read, per 1M tokens	$0.01	$0.02
Cache write, per 1M tokens	$0.125	$0.25
Context window	1,000,000 tokens	922,000 tokens
Max output	128,000 tokens	128,000 tokens
Image input	Yes	Yes
Tool calling	Yes	Yes
Structured output (JSON)	Yes	Yes
Prompt caching	Yes	Yes
Reasoning	Yes	Yes
Providers on LiteLLM	7	9
Price shown is from	Anthropic	OpenAI
Added to LiteLLM	–	July 9, 2026

Cheapest provider for each

Claude Haiku 5.5: Anthropic at $0.10 input and $0.50 output per 1M tokens (claude-haiku-5-5)
GPT-5.6 Luna: OpenAI at $0.20 input and $1.20 output per 1M tokens (gpt-5.6-luna)

Benchmark scores

From the Epoch AI Benchmarking Hub (CC BY 4.0), data as of October 9, 2026. Some results were run by Epoch and others collected by Epoch from each benchmark's own leaderboard, linked in the first column. Each model's best reported reasoning setting is shown in brackets. Higher is better. Only benchmarks with a score for both models are shown.
Benchmark	Claude Haiku 5.5	GPT-5.6 Luna
Overall
Epoch Capabilities Index	159.3	156.5
Coding and agents
FrontierCode	46.4% (max)	39.8%
APEX-Agents	56.1% (max)	43.0%
Reasoning and knowledge
GPQA Diamond	89.6% (max)	91.6% (max)
ARC-AGI-2	61.3% (xhigh)	59.5% (max)
SimpleQA Verified	23.8% (max)	41.0% (max)
Math
FrontierMath (tiers 1-3)	75.1% (max)	82.1% (max)
FrontierMath (tier 4)	46.3% (max)	61.0% (max)
OTIS Mock AIME 2024-2025	98.9% (xhigh)	98.3% (max)

When to pick which

Pick Claude Haiku 5.5 for

Costs 50% less per input token: $0.10 vs $0.20 per 1M.

From differences in the data above only; where the data is equal, no reason is given.

Call Claude Haiku 5.5 or GPT-5.6 Luna with LiteLLM

Both use the OpenAI format through LiteLLM, so switching is a model name change. Set each provider's API key in the environment first.

Python SDK

from litellm import completion

messages = [{"role": "user", "content": "Hello!"}]

# Claude Haiku 5.5
response = completion(model="claude-haiku-5-5", messages=messages)

# GPT-5.6 Luna
response = completion(model="gpt-5.6-luna", messages=messages)

# Claude Haiku 5.5, falling back to GPT-5.6 Luna if the call fails
response = completion(model="claude-haiku-5-5", messages=messages, fallbacks=["gpt-5.6-luna"])

FAQ

Which is cheaper, Claude Haiku 5.5 or GPT-5.6 Luna?

Claude Haiku 5.5 is cheaper at 3 input tokens per output token: $0.20 per 1M tokens blended ($0.10 input, $0.50 output), versus $0.45 ($0.20 and $1.20) for GPT-5.6 Luna, as of October 21, 2026.

Which has the larger context window, Claude Haiku 5.5 or GPT-5.6 Luna?

Claude Haiku 5.5 has the larger context window: 1,000,000 tokens, versus 922,000 for GPT-5.6 Luna.

Can I switch from Claude Haiku 5.5 to GPT-5.6 Luna without changing my code?

Yes. Through LiteLLM both use the OpenAI format, so switching means changing the model name from "claude-haiku-5-5" to "gpt-5.6-luna", or pointing a LiteLLM gateway model alias at the other model. You can also use one as a fallback for the other.