OpenAI releases GPT-6 Sol and Luna and slashes token prices
According to The New Stack, OpenAI added GPT-6 Sol and Luna on Tuesday next to flagship GPT-6 Astra, with reporting noting no GPT-6 Terra. The New Stack lists Sol at $2/$10 and Luna at $0.10/$0.50 per million input and output tokens, at least fifty percent below GPT-5.6 Sol and Luna figures in its coverage. An OpenAI spokesperson told The New Stack these are default prices, not promotions, linking lower serve cost to caching and inference gains for users. TechCrunch portrayed both as cut from Astra’s cloth with cheaper inference positioning. Teams budgeting API spend may gain if defaults stick, yet savings still ride on workload match and unsettled benchmarks. The New Stack flagged missing full benchmarks at launch, so GPT-5.6 capability deltas remain tentative.
OpenAI releases GPT-6 Sol and Luna and slashes token prices
The New Stack reports OpenAI launched GPT-6 Sol and Luna on Tuesday to sit alongside flagship GPT-6 Astra, with no GPT-6 Terra in the lineup yet. It also cites per-million input and output token prices cut by half or more versus the prior generation, including GPT-6 Sol at $2/$10 per million tokens.
Key takeaway
Treat the launch as a default token-price reset (roughly half GPT-5.6 list rates) more than a proven leap in capability until full benchmarks land.
What happened
The New Stack reports OpenAI launched GPT-6 Sol and Luna on Tuesday alongside flagship GPT-6 Astra, with no GPT-6 Terra announced yet in the lineup.
Reporting cites Sol at $2/$10 and Luna at $0.10/$0.50 per million input/output tokens, cuts of half or more versus GPT-5.6 Sol and Luna pricing, with an OpenAI spokesperson calling the new rates default rather than promotional.
Evidence
OpenAI released GPT-6 Sol and Luna to complement GPT-6 Astra and has not announced GPT-6 Terra.
The New Stack AI · attributed
OpenAI on Tuesday released GPT-6 Sol and Luna , which will complement the flagship GPT-6 Astra model in OpenAI’s lineup. As of now, there is no GPT-6 Terra.
GPT-6 Sol and Luna token list prices were cut by half or more versus GPT-5.6 Sol and Luna.
The New Stack AI · attributed
GPT-6 Sol will cost $2/$10 per million input/output tokens (vs. $4/$20 for GPT-5.6 Sol ), and Luna will come in at $0.10/$0.50 (vs. $0.20/$1.20).
OpenAI framed the new GPT-6 rates as default pricing tied to caching and inference efficiency.
The New Stack AI · attributed
for the new GPT-6 models, this is the default price, an OpenAI spokesperson tells The New Stack
TechCrunch reported two new OpenAI models positioned alongside Astra with a lower-cost angle.
TechCrunch AI · attributed
OpenAI is launching two new models, which it says are cut from the same cloth as Astra.
The New Stack did not receive a full benchmark set from OpenAI ahead of launch.
The New Stack AI · attributed
Unlike with previous releases, OpenAI did not provide us with a full set of benchmark results ahead of the launch.
Why it matters
If default list prices hold, high-volume inference workloads could see lower per-token spend, but realized savings still depend on workload fit and unverified performance data.
Limits and uncertainties
The New Stack states OpenAI did not provide a full benchmark set before launch, so capability comparisons to GPT-5.6 remain provisional.
The packet does not establish GPT-6 Terra timing, full Luna/Sol benchmark tables, or independent confirmation that efficiency gains fully explain the price cuts.
Practical implications
Operators should reprice API budgets against Sol at $2/$10 and Luna at $0.10/$0.50 per million tokens while treating capability gains as unquantified until benchmarks arrive.
Builders comparing GPT-5.6 to GPT-6 Sol/Luna should delay production switches that depend on benchmark-sensitive quality until The New Stack or OpenAI publishes complete results.
What to watch
Publication of OpenAI’s full benchmark package and The New Stack’s promised story update once results are supplied.
Whether OpenAI announces GPT-6 Terra or changes default GPT-6 Sol/Luna list pricing after launch.
