MODEL PROFILE / EDITION 2026.09
GPT-5.6 Luna
Cost-sensitive teams running lots of code edits and CI-style fixes
Balanced model evidence score; rank 25. No rank delta is shown because there is no prior numeric snapshot.
Best for
Cost-sensitive teams running lots of code edits and CI-style fixes
Price & access
$0.20 in / $1.20 out per 1M; cached input $0.02
Evidence summary
First-party pricing/context plus third-party SWE-bench Pro; frontend and agentic evidence is thin and partly family-level.
Edition status
No prior snapshot. Tracked in the September 2026 baseline.
Strengths
- SWE-bench Pro 62.7% at $0.20/$1.20 is an extreme price/quality ratio
- 1.05M context at budget pricing
- 127.2 tok/s throughput
- Cached input $0.02 makes repeat-context agents very cheap
Tradeoffs
- Arena frontend Elo 1516 is near the bottom of this cohort
- TTFT 124.70 s at max effort
- AA Intelligence Index 43 is well below the frontier
- No visual/design benchmark evidence
Score profile
visual 62 · code 76 · agentic 70 · debug 68 · context 86 · speed 66 · value 93
Linked evidence
Evidence source 1 ↗Evidence source 2 ↗Evidence source 3 ↗Evidence source 4 ↗Evidence source 5 ↗Compare this model on the full scoreboard
Open comparison view →Use the same published data, filters, and side-by-side metrics.