GUIDE

GPT-5.6 Luna Long-Context Recall Test (~300k)

GPT-5.6 Luna advertises a ~1.05M context window. Secondary writeups claim recall collapses near ~300k. We ran a lean, repeatable needle-in-haystack suite on our Hetzner OpenAI path — and publish the numbers, including what we could not reproduce.

8–10 min read

Key takeaways
  • Measured 2026-07-31 on Hetzner via OpenAI Chat Completions (reasoning_effort: none, temperature 0)
  • Phase 1: Luna exact-cited a mid-document needle at ~8k → ~299,112 prompt tokens; Terra also passed ~300k mid-depth
  • Phase 2: Luna passed depths 0.1 / 0.5 / 0.9 at 128k, 200k, and 300k, plus summarize-then-cite at 300k (11/11 scored cells correct)
  • Viral “breaks at 300k” claim: not reproduced on this synthetic NIAH harness
  • Limits: one needle, synthetic filler, short answers — not multi-hop RAG, not adversarial distraction, not a full 1M sweep

Sibling: Luna guide · Sol vs Terra vs Luna · cost per completed task.

Method (repeatable, cheap)

We plant a unique CLOUDYBOT_NEEDLE_… string inside a long synthetic inventory document, then ask the model to return only that code. Prompt size is taken from API usage.prompt_tokens. Script: sites/cloudaxis/content/gpt-5-6-luna-long-context-recall.mjs; logged results in gpt-5-6-luna-long-context-measurement.mjs.

Results

Phase 1 — mid-depth Luna ladder: exact cite at every checkpoint through 299,112 prompt tokens (3734 ms). Terra mid-depth at ~300k also exact-cited (299,112 tokens).

Model Target Prompt tokens Result
gpt-5.6-luna~8k7,798exact cite
gpt-5.6-luna~32k31,104exact cite
gpt-5.6-luna~64k63,110exact cite
gpt-5.6-luna~128k127,122exact cite
gpt-5.6-luna~200k199,116exact cite
gpt-5.6-luna~300k299,112exact cite
gpt-5.6-terra~128k127,122exact cite
gpt-5.6-terra~300k299,112exact cite

Phase 2 — depth sweep: Luna at 128k / 200k / 300k × depths 0.1, 0.5, 0.9 — all exact cites. Summarize-then-cite at 300k mid-depth: exact cite. Terra early-depth (0.1) at 300k: exact cite. Hard-sweep failures: 0.

What this does and does not prove

It proves that on this NIAH harness, Luna still exact-cites a single planted string through ~300k prompt tokens across early/mid/late depths — including after a forced short summary. It does not prove perfect usable recall across a full 1M window, multi-hop retrieval, or noisy real corpora.

Reproduce

On a host with OPENAI_API_KEY (we used the CloudyBot API container on Hetzner):

node sites/cloudaxis/content/gpt-5-6-luna-long-context-recall.mjs
# optional harder sweep: e1-hard-sweep.mjs (depth × summarize)

Frequently asked questions

Does GPT-5.6 Luna really fail long-context recall at 300k?

Not on our needle-in-haystack harness dated in this post. Luna exact-cited through ~299k prompt tokens at early, mid, and late depths, and after a summarize-then-cite prompt. Other tasks may still degrade earlier.

Did you test the full 1M window?

No. This pass stops near 300k to control cost. The methodology script can be extended.

How is this different from the context window spec?

Context window is capacity. Recall is whether the model can use a fact buried in that capacity. Specs do not equal usable recall on every task.

Can I reproduce this?

Yes. The measurement scripts and JSON logs ship under sites/cloudaxis/content/. You need an OpenAI API key and the gpt-5.6-luna model id.

Related reading in this series
Luna complete guide · Luna landing · Cost per completed task