gpt-5.4-mini vs gpt-6-luna
Fourteen priced metrics for the same 2,000-in / 300-out call, with and without provider caching, and what TokenZero's cascade would do with the pair.
At list prices as of 2026-10-26, gpt-6-luna costs $0.15 per million blended tokens on a 2,000-in / 300-out call against $1.24 for gpt-5.4-mini, 87.7% less. With the static prefix cached both drop further: $1.84 versus $0.22 per thousand calls. TokenZero routes only prompts its classifier scores simple and holds a 99.8% parity floor.
· Source: provider list prices verified in pricing table 2026-10-26; costs computed by the same code that signs TokenZero receipts
At a glance
- Cheaper per blended token
- gpt-6-luna by 87.7% (computed)
- gpt-5.4-mini blended $/1M
- $1.24 (computed)
- gpt-6-luna blended $/1M
- $0.15 (computed)
- Input $/1M
- $0.75 vs $0.10
- Output $/1M
- $4.50 vs $0.50
- Cached input $/1M
- $0.075 vs $0.010
- Cache discount
- 90% vs 90% (computed)
- Monthly at 100k calls
- $285 vs $35 (computed)
- Saved per 1M calls
- $2,500 (computed)
- Providers
- openai vs openai
- Lifecycle
- active, no retirement announced vs active, no retirement announced
- Parity floor for routing
- 99.8% (shadow-eval breaker)
gpt-5.4-mini vs gpt-6-luna: list prices and the same workload on both (USD, table 2026-10-26)
| Metric | gpt-5.4-mini | gpt-6-luna | Delta |
|---|---|---|---|
| Provider · tier | openai · fast | openai · fast | same provider |
| Input $/1M | $0.75 | $0.10 | 86.7% gap |
| Output $/1M | $4.50 | $0.50 | 88.9% gap |
| Cached input $/1M | $0.075 | $0.010 | 90% / 90% off input |
| Cache write $/1M | no surcharge | $0.125 | |
| Minimum cacheable prefix | 1024 tokens | 1024 tokens | |
| Effective blended $/1M (2,000 in / 300 out) | $1.24 | $0.15 | gpt-6-luna 87.7% less |
| Per call | $0.00285 | $0.00035 | $0.00250 |
| Per month at 100k calls | $285 | $35 | $250 |
| …with the 1,500-token prefix cached | $184 | $22 | $162 |
| Long-context threshold | single tier | 272,000 tokens ($0.20 in / $0.75 out above) | |
| Processing tiers | flex 0.5x · fast 2x | flex 0.5x · fast 2x | |
| Lifecycle | active, no retirement announced | active, no retirement announced | |
| Cheaper sibling (cascade target) | — | — | |
| TokenZero cascade recommendation | gpt-5.4-mini → gpt-6-luna | $2,500 saved per 1M calls if every call moved; $1,000 at a 40% simple-prompt share | 99.8% parity floor, breaker on |
Use either model through TokenZero
OPENAI_BASE_URL=https://tokenzero.sh/v1
# keep your own keys; cache formatting on, routing advisory, signed receipt per callFree up to $500/month of routed spend; paid plans take 15% of verified savings with a 3x-ROI-or-$0 rule.
What is not on this page
Third-party quality scores (coding ELO, extraction accuracy, reasoning benchmarks) and per-model latency are not shown because TokenZero does not measure them; the numbers above are prices and the arithmetic on them. The latency benchmark shows what the gateway does measure.
Questions people ask
Which is cheaper, gpt-5.4-mini or gpt-6-luna?
gpt-6-luna. At list prices as of 2026-10-26 it costs $0.15 per million blended tokens versus $1.24 for gpt-5.4-mini, a 87.7% gap on a 2,000-input / 300-output call. Per month at 100,000 calls that is $35 versus $285.
How much does prompt caching change the gpt-5.4-mini vs gpt-6-luna gap?
Cached input costs $0.075 on gpt-5.4-mini and $0.010 on gpt-6-luna per million. With a 1,500-token static prefix cached, 100,000 calls cost $184 and $22 respectively; TokenZero's formatter places the breakpoints and sets prompt_cache_key automatically.
Can TokenZero route gpt-5.4-mini traffic to gpt-6-luna safely?
Yes, within openai: with X-TokenZero-Cascade: auto the classifier scores each prompt and only simple ones move; shadow evaluation re-runs a sample on gpt-5.4-mini and a pair below 99.8% parity is auto-disabled. Moving every call would save $2,500 per million calls; at a 40% simple share it is $1,000.
Is gpt-5.4-mini or gpt-6-luna being retired?
gpt-5.4-mini: active, no retirement announced. gpt-6-luna: active, no retirement announced. TokenZero audits flag retired ids and the gateway warns in X-TokenZero-Warning with the provider's replacement.