OpenAI has introduced two new models, GPT-6 Sol and GPT-6 Luna, which cost half as much as their predecessors while delivering similar performance levels. The company states these updates aim to compete with pricier options from Anthropic, specifically targeting cost efficiency rather than raw capability gains.
The pricing drop is direct. GPT-6 Sol now charges $2 for every million input tokens and $10 for output. Luna is cheaper still, at $0.10 for input and $0.50 for output. This brings OpenAI’s rates into the same bracket as cheaper open-weight models. Terra, previously the cheapest option in the range, is no longer available.
Pricing breakdown
- GPT-6 Sol input: $4 per million down to $2
- GPT-6 Sol output: $20 per million down to $10
- GPT-6 Luna input: $0.20 per million down to $0.10
- GPT-6 Luna output: $1.20 per million down to $0.50
OpenAI credits improvements in caching and inference for the reduction, claiming it passes savings directly to users. The new models target specific workflows. Sol handles recurring complex tasks like feature building, code review, debugging, and data analysis. Luna manages large volumes of well-defined tasks such as document summarisation and information extraction at a low cost.
Caching and availability
Prompt caching has improved, offering a 90 percent discount on cached input tokens. A new dashboard and diagnostics tool help developers optimise usage. Users can now adjust reasoning effort and tool availability without invalidating the cache.
Both models launch in ChatGPT Work and Codex for Plus, Pro, Business, Enterprise, and Edu subscribers. Free and Go users access Luna via the desktop app. Neither model is available in regular chat initially. The API lists them as
gpt-6-sol
and
gpt-6-luna
, while ChatGPT access rolls out gradually.
Performance claims versus cost
OpenAI positions the new models around price-to-performance, comparing them against Anthropic’s Claude lineup. On OSWorld 2.0, which tests computer use, GPT-6 delivers results similar to Claude Opus 5 at roughly 80 percent lower cost, though Astra still leads in this category.
On AutomationBench, which tests business workflows across 47 tools, GPT-6 Sol at its highest effort level reportedly beats Claude Opus 5 at maximum effort. OpenAI puts Sol’s cost per task at just 9 percent of Opus’s. Luna improves on its predecessor by 5.4 percentage points while costing 58 percent less.
For coding, OpenAI cites two benchmarks. FrontierCode 1.1 checks whether AI agents produce code integratable into an existing codebase, verifying test quality, style, and compliance. GPT-6 Sol scores 49.3 percent at maximum effort for $2.14 per task. This puts it roughly on par with Claude Fable 5.1, which scores 50.3 percent at maximum effort but costs six times as much at $12.83. Claude Opus 5 reaches 53.4 percent for $4.31 at medium effort, the setting that produced its best result on this test.
Software engineering benchmarks
On DeepSWE v1.1, a benchmark for demanding software engineering tasks over long stretches in real codebases, OpenAI reports 68.8 percent for GPT-6 Sol at maximum effort. That is within 1.1 percentage points of Claude Fable 5’s best score of 69.9 percent at “xhigh,” while Fable at “max” hits 69.7 percent for $21.63 per task.
Sol is not chasing the frontier here. Claude Opus 5 reaches 73.7 percent at maximum effort for $11.84 per task, and OpenAI’s own GPT-5.6 Sol scores 72.7 percent for $6.46. The point of GPT-6 Sol is to land close to those numbers for a fraction of the price. At “xhigh,” it delivers 66.6 percent for $1.00 per task. Luna is cheaper still, matching that score at maximum effort for just $0.22. OpenAI says Luna’s result is comparable to Claude Opus 5 and Claude Fable 5 at medium effort, while costing 93 percent less than Opus and 96 percent less than Fable.
The DeepSWE results also complicate the choice between OpenAI’s own models. Luna at maximum effort matches Sol at “xhigh” while costing 78 percent less. Cranking Sol up to maximum effort only gets you to 68.8 percent, a mere 2.2 percentage points above Luna. It is unclear who should sort through all this in real-world use.
Overall, OpenAI’s benchmark selection looks cherry-picked. Metrics like GDPval for knowledge work or Terminal-Bench 4.0 for agentic coding are missing, even though they are part of the usual lineup. OpenAI also appears to have missed the launch of Opus 5.5, which is potentially up to 40 percent cheaper than Opus 5 with significantly better performance.
Independent analysis sees little actual progress
According to Artificial Analysis, GPT-6 Sol and Luna cut per-task costs in half compared to their predecessors, but intelligence scores stay at GPT-5.6 levels, with gains in some evaluations and regressions in others. On the coding agent index, Sol improves by 2 points while Luna drops by 2, according to the analysis.
Artificial Analysis also found regressions on two key knowledge-work benchmarks. On GDPval-AA v2.1, which tests computer-based knowledge work across 44 professional fields, Sol loses about 100 Elo points and Luna drops about 75. Manual inspection traced the regressions mainly to lower presentation quality and incomplete results.
Artificial Analysis has faced criticism before, though, when it rated OpenAI’s Astra model too low because of outdated benchmarks. Two benchmark updates followed, after which Astra was back on top. We will see what happens this time.
Either way, it is very much clear that OpenAI is betting heavily on price with GPT-6 Sol and Luna. Now the models need to prove themselves in day-to-day work, since benchmarks only tell part of the story. That may also explain why OpenAI left some of them out. Combined with the different reasoning levels and results that sometimes look tuned for benchmarks, whether intentionally or not, the whole benchmarking game seems more ridiculous with each new model launch.
