GPT-5.6 Luna

Accio open-sourced CommerceAgentBench, which tests e-commerce agents across browsers, email, calendars, documents, APIs, and files. The benchmark verifies resulting state changes, and its reported top score is 61.7%.

Together says two V4 Flash attempts solved more DeepSWE tasks than one GPT-5.6 Luna attempt for roughly one-third the cost. Practitioners report Flash-0731 results vary sharply by harness and pass count.

Practitioners reported better Codex multi-agent runs by raising concurrency and splitting work across Sol, Terra, and Luna. One workflow sends deploy tasks to Luna Max to preserve Sol tokens.

Sol-advisor routes Codex tasks through GPT-5.6 Sol, Luna, and Terra, while users compare Luna Max as a lower-cost reasoning setting. Early reports say small routing tests need larger benchmarks.

New tests showed DeepSeek V4 Flash as cheaper per token and faster on some serving paths. Ramp said it cost 3x more than GPT-5.6 Luna per SWE-Bench task because it used more turns.

OpenAI said GPT-5.6 Luna pricing fell 80%, while Terra fell 20%. Codex users recommended max reasoning for linting, tests, and dependency work, but cautioned against forcing Luna into subagent roles.

OpenAI said GPT-5.6 Luna is 80% cheaper and Terra is 20% cheaper, with lower usage burn in Codex and ChatGPT Work. Sol Fast adds up to 2.5x speed at 2x price, and gateways reflected the new pricing.

OpenAI posts said GPT-5.6 Sol helped post-train GPT-5.6 Luna, framing Sol as a research agent rather than just a coding model. Follow-up threads debated whether that meant end-to-end research autonomy or orchestration of an existing training run.