Can you trust Claude Haiku 5.5 for space work?
By Anthony Vincent Caracappa, Founder and Self-Directed Space Researcher, Viventine Space Systems · updated
On its own, trust it for baseline knowledge. Connected to Orbit Sentinel, trust it for catalog lookup, debris & conjunction context, regulatory & licensing, company & market intelligence, baseline knowledge.
| Job | AloneClosed-book | With dataWith Orbit Sentinel tools | n | ||||
|---|---|---|---|---|---|---|---|
| right / wrong | range | declined | right / wrong | range | declined | ||
| Catalog lookup | 27 / 24 | 16–41 | 49 | 100 / 0 | 92–100 | 0 | 45×3 |
| Constellation & fleet status | 26 / 16 | 15–40 | 58 | 87 / 12 | 74–94 | 2 | 43×3 |
| Debris & conjunction context | 19 / 10 | 10–33 | 71 | 93 / 0 | 81–98 | 7 | 42×3 |
| Regulatory & licensing | 23 / 8 | 13–38 | 69 | 82 / 3 | 67–91 | 16 | 38×3 |
| Company & market intelligence | 27 / 10 | 16–42 | 63 | 94 / 2 | 83–98 | 3 | 41×3 |
| Ground infrastructure | 13 / 3 | 5–26 | 85 | 89 / 6 | 76–96 | 5 | 40×3 |
| Baseline knowledge | 100 / 0 | 87–100 | 0 | 97 / 0 | 82–100 | 3 | 25×3 |
| Orbit Sentinel-only questions | 0 / 0 | 0–13 | 100 | 100 / 0 | 87–100 | 0 | 25×3 |
% right / % wrong per cell; color is the verdict (green trust it, amber verify, red don't use). Range is the 95% interval for % right, counting each question once; n is questions × runs for the tools track.
- How it was run
- Claude Code on a Claude Pro subscription
- Tool loop
- Claude Code's own tool loop
- Runs
- 1 on its own, 3 with tools; rates pool all runs, ranges count each question once
- Verdict rule
- Trust it: wrong ≤ 5% and right ≥ 80%. Don't use it: wrong > 20%. Otherwise verify its answers.
Limits, harness differences and corrections: methodology.