OpenAI Releases GPT-5.6 Luna Pro with Extended Reasoning Mode at $1/$6 Per Million Tokens
OpenAI has released GPT-5.6 Luna Pro, a reasoning-enhanced variant of GPT-5.6 Luna with a 1 million token context window. The model is priced at $1 per million input tokens and $6 per million output tokens, with a knowledge cutoff date of February 2026.
GPT-5.6 Luna Pro — Quick Specs
OpenAI Releases GPT-5.6 Luna Pro with Extended Reasoning Mode
OpenAI has released GPT-5.6 Luna Pro, a new variant of its GPT-5.6 Luna model configured with extended reasoning capabilities for complex tasks. The model features a 1 million token context window and is priced at $1 per million input tokens and $6 per million output tokens.
Technical Specifications
GPT-5.6 Luna Pro uses the same underlying architecture as the standard GPT-5.6 Luna model but operates with reasoning.mode set to "pro" for higher-quality responses on complex tasks. The model's knowledge cutoff date is February 2026, with an official release date of July 9, 2026.
The model supports multimodal inputs and outputs, though specific modality details were not disclosed in the release.
Pricing and Availability
The model is currently hosted by a single provider on OpenRouter, which forwards all requests directly without routing decisions. According to OpenRouter data, effective pricing can be 60-80% lower than list prices when prompt caching is utilized with repeated context.
Pricing structure:
- Input: $1 per 1 million tokens
- Output: $6 per 1 million tokens
- Context window: 1 million tokens
Performance Characteristics
OpenRouter monitors throughput (tokens per second), latency (total round-trip time), and time-to-first-token (TTFT) for all models on its platform. The company also tracks uptime as the percentage of successful requests over 30-day rolling windows, with automatic failover to alternative providers when errors occur.
What This Means
The "pro" reasoning mode designation suggests OpenAI is segmenting its model offerings by computational intensity rather than just parameter count or architecture. The $6 output pricing is notably higher than the $1 input cost, indicating the reasoning mode requires significantly more compute resources during generation. The 1M context window positions this model for long-document analysis and complex reasoning tasks, though the lack of benchmark scores makes direct capability comparisons difficult. The February 2026 knowledge cutoff and July 2026 release date indicate this is forward-looking information rather than a current release.
Related Articles
AWS Benchmark: OpenAI's GPT-5.6 Luna Beats GPT-5.4 Mini on Cost-Per-Correct-Answer Despite Similar List Price
An AWS blog post using an open-source benchmarking harness finds that GPT-5.6 Luna, Terra, and Sol on Amazon Bedrock deliver lower cost-per-correct-answer than OpenAI's cost-optimized GPT-5.4 Mini and Nano, once accuracy, token efficiency, and agent turn counts are factored in. The analysis also cites a July 30, 2026 price cut of up to 80% for GPT-5.6 Luna on Amazon Bedrock.
OpenAI's GPT-6 Astra Beats Claude Fable 5.1 Nearly 3-to-1 in Autonomous Business Benchmark, Tops Drone Navigation Tests
Independent testing lab Andon Labs found OpenAI's GPT-6 Astra nearly triples Claude Fable 5.1's performance running a simulated vending machine business, averaging $15,515 versus $5,422. Astra also became the first model to beat human-AI baseline performance across all five Drone-Bench subtasks, including autonomous person-tracking via drone.
GPT-6 Astra Beats Ai2's MolmoAct2 on New Robotics Benchmark, Researcher Calls It a 'Step Change'
A new robotics benchmark called StationeryBench shows OpenAI's GPT-6 Astra completing 7 of 100 desk-object manipulation tasks versus zero for Ai2's MolmoAct2, with a median progress score of 46 against 12. Cornell/DeepMind researcher Yoav Artzi calls the result a 'step change in spatial reasoning.'
Perplexity Says It Runs End-to-End Engineering Systems on OpenAI's GPT-6 Astra
Perplexity says it has shifted core engineering workflows, including code changes and production monitoring, onto OpenAI's GPT-6 Astra model. The claim comes from an OpenAI-published case study with no independent benchmark data released.
Comments
Loading...