GPT 5.6 Sol

On June 26, 2026, OpenAI began a limited preview of the GPT-5.6 series, headlined by GPT-5.6 Sol as its flagship model, alongside Terra (balanced, positioned as competitive with GPT-5.5 at roughly 2x cheaper) and Luna (fast and affordable, lowest cost). The page emphasizes that Sol launches with OpenAI's most robust sa OpenAI described Sol as its strongest model yet, sharing preview evaluations highlighting improved agentic capabilities in coding, biology, and cybersecurity, with additional safety and preparedness evaluations available in the system card. Preview access was coordinated with the U.S. government under a trusted-partner

The llm-stats tracker page for GPT-5.6 Sol reports a composite LLM Stats Score of 55.0 at a blended cost of approximately $6.19 per million tokens, placing Sol ahead of GPT OSS 120B and behind GPT-6 Astra on the leaderboard. Conversation-depth telemetry shows Sol slightly improving through turns 2–10 (+0.4) before decl The page also lists internal OpenAI benchmarks evaluated at max reasoning effort, including a 1.00 pass rate on the Connectors production benchmark and 0.97 on Capture-the-Flag Challenges, as well as category-specific quality movements in Chat (+2.01σ), Websites (−1.22σ), SVG, Games, Animations, 3D, Audio, and Data Viz

CodeRabbit's independent coding-focused evaluation of GPT-5.6 Sol and Terra concludes that Sol is the model to test for long-horizon coding work, where it shows stronger follow-through across multi-file tasks, tests, and follow-up fixes. In CodeRabbit's harness, Sol surfaced more review issues than prior OpenAI models, The assessment frames Sol as a practical rather than purely intellectual improvement: it does not displace Fable 5 on architectural judgment or Sonnet 5 on comment quality, but wins on the more useful axis of finishing more assigned work. The piece advises engineering teams to measure cost per solved task alongside per

BenchLM's tracking page for GPT-5.6 Sol (released July 9, 2026) reports a composite capability score of 79.6/100, ranking the model 5th out of 232 tracked models, with category percentiles of 97th in Agentic, 98th in Coding, 38th in Reasoning, 98th in Knowledge, and 94th in Multimodal. Its strongest eligible category i API pricing is listed at $5 per million input tokens and $30 per million output tokens, with cached input at $0.50 and a blended effective rate of $17.50. Context window is 1.05M tokens and throughput is reported at 77 tok/s with a first-token latency of 107.13s. As of September 4, 2026 data, the page flags GPT-5.6 Sol

OpenAI announced the public release of its GPT-5.6 model family on Thursday, July 9, 2026, roughly two weeks after initially limiting access to a small group of trusted partners at the U.S. government's request. The release includes GPT-5.6 Sol (the flagship), plus Terra and Luna variants. CEO Sam Altman confirmed the Alongside the GPT-5.6 rollout, OpenAI introduced GPT-Live, a new generation of voice models capable of listening and speaking simultaneously. GPT-Live-1 and GPT-Live-1 mini began rolling out to ChatGPT users worldwide on Wednesday. The launch ends the government's vetted-partner-only period; OpenAI stated it "doesn't b

Nextgov/FCW reported on July 8, 2026 that OpenAI's GPT-5.6 series, including Sol, Terra, and Luna, would become publicly available later that week, with Thursday's launch announced via a post on X. GPT-5.6 Sol is described as the strongest model in the series, tuned for biology, chemistry, and cybersecurity workloads, The article notes that delays to the public rollout were attributed to the powerful capabilities of the models, and that OpenAI did not clarify whether changes (to parameters, model weights, or architecture) were made prior to release. The coverage frames the expansion as OpenAI "expanding preview access globally" afte

A Towards AI write-up from late June 2026 walks through the GPT-5.6 family launch, emphasizing that the Sol, Terra, and Luna naming scheme is durable, meaning each tier can advance on its own cadence independent of the generation number. Pricing is confirmed at $5/$30 per million tokens for Sol (matching GPT-5.5), $2.5 The article highlights product-side additions including a max reasoning effort on Sol, an Ultra mode that uses subagents, a 30-minute prompt cache, and Cerebras deployment. It also notes that the OpenAI system card itself flags areas where the company remains uncertain about model behavior, providing a more grounded vi

Explainx.ai detailed the July 9, 2026 rollout of GPT-5.6 Sol, Terra, and Luna across ChatGPT, Codex, and the API, with published GA benchmarks including Agents' Last Exam (Sol 53.6, +13.1 adaptive), Artificial Analysis Coding Agent Index (80.0, +2.8 at less than half the tokens and one-third the cost), and Terminal-Ben The explainer also surfaces product details beyond raw benchmarks: Sol supports max reasoning effort, with Ultra mode using subagents for parallel multi-agent coordination; a 30-minute prompt cache is referenced; and Cerebras is listed as a deployment partner offering Sol at up to 750 tokens per second for select custo