OpenAI launches GPT-6.1 Terra at half the price of GPT-5.6 Terra
Current overview (not historical)
OpenAI announced GPT-6.1 Terra on October 26, 2026, a mid-tier model positioned between its smaller and larger offerings. On the same day, the company halved API pricing compared to GPT-5.6 Terra, setting input token costs at $1.00 per million and output at $5.00 per million, with cached input at $0.05. The model is now available to ChatGPT Plus, Pro, Enterprise, and Edu users within ChatGPT Work and Codex.
Written with AI from the reporting · updated 10/29/2026
Expand the detailed fact route and sources
Story route
A station-by-station route from the first public evidence onward; every station carries its own evidence and uncertainty.
- 1Unclassified fact stationOct 26, 00:00recorded Oct 27, 03:24revised
OpenAI发布GPT-6.1 Terra
Read original source ↗ The DecoderReplay this moment in the explorerEvidence & revisions · 6 source records
- Subject：
- openai
- Action：
- 发布
- Object：
- GPT-6.1 Terra
- Occurred (fact side)：
- Oct 26, 00:00
- Published (valid side)：
- Oct 27, 04:30
- First seen (known side)：
- Oct 27, 05:26
- Published (valid side)：
- Oct 26, 08:00
- First seen (known side)：
- Oct 27, 03:24
- Published (valid side)：
- Oct 27, 05:46
- First seen (known side)：
- Oct 27, 06:14
- Published (valid side)：
- Oct 27, 07:53
- First seen (known side)：
- Oct 27, 08:28
- Published (valid side)：
- Oct 27, 17:48
- First seen (known side)：
- Oct 27, 18:04
- Published (valid side)：
- Oct 27, 18:09
- First seen (known side)：
- Oct 27, 18:39
Fact stationRevisedCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
- 2Unclassified fact stationOct 28, 00:00recorded Oct 29, 08:15single sourcerevised
Ted Sanders 解释 OpenAI 模型命名规范
Read original source ↗ Simon WillisonReplay this moment in the explorerEvidence & revisions · 1 source records
- Subject：
- Ted Sanders
- Action：
- 解释模型命名规范
- Object：
- OpenAI 的模型命名体系
- Occurred (fact side)：
- Oct 28, 00:00
- Published (valid side)：
- Oct 29, 07:23
- First seen (known side)：
- Oct 29, 08:15
Fact stationSingle source · one-sided claimRevisedCopy may have been revised later; time replay is bounded. Layer labels group reading only, not causation between layers.
Uncertainty marks flag where to check the sources; they are not proven causation.
Report timeline
Explore every side through dated source reports.
- Simon WillisonRelated fact: single sourceSource report revisedOpenAI Explains How Model Size Drives Performance and Naming Conventions
Report first recorded Oct 29, 08:15
Ted Sanders clarified that OpenAI’s large language models follow a performance hierarchy—Luna < Terra < Sol < Astra—where larger models deliver better accuracy, higher reliability, and richer recall of obscure facts, but at a higher per‑token cost. He explained that version numbers such as 6.1 indicate release order, shared training data or personality traits with other 6.1 models, and a moderate improvement rather than a major leap like a 7.x release.
Read original source ↗
- The DecoderSource report revisedOpenAI's Ted Sanders explains GPT-6.1 Terra naming and model performance factors
Report first recorded Oct 27, 18:39
OpenAI's Ted Sanders categorizes the new GPT-6.1 Terra, stating that model performance depends on multiple dimensions including size, reasoning effort, and multi-agent systems. He notes that newer, smaller models can outperform larger ones through better training data or architecture. GPT-6.1 Terra is priced lower than Astra and performs similarly in image processing and computer use, but falls short in most other areas.
Read original source ↗ - BeInCryptoOpenAI launches GPT-6.1 Terra with 50% price cut, underperforms Anthropic models in independent tests
Report first recorded Oct 27, 18:04
OpenAI released GPT-6.1 Terra on October 26, reducing input token price to $1 and output token price to $5, half the cost of its predecessor GPT-5.6 Terra. Independent tester Artificial Analysis scored Terra 48 on its Intelligence Index, ranking it below Anthropic's Claude Opus 5.5 (58 points) and Claude Sonnet 5.5 (56 points). Despite the lower score, OpenAI emphasizes Terra's cost efficiency, stating it costs about 9 times less per task than Opus 5.5 on Agents' Last Exam, and positions the model for high-volume tasks like invoice processing.
Read original source ↗ - Simon WillisonSource report revisedOpenAI launches GPT-6.1 Terra with visual reasoning and computer use capabilities
Report first recorded Oct 27, 08:28
OpenAI released GPT-6.1 Terra on October 26, 2026, as a mid-range model positioned between Sonnet and Opus, emphasizing visual reasoning and computer use for tasks like drafting proposals and editing videos. Priced at $1 per million input tokens and $5 per million output, it costs nearly half as much as GPT-5.6 Terra and GPT-6.1 Sol, while being around 10 times more expensive than GPT-6 Luna. OpenAI's Ted Sanders noted on Hacker News that Terra is better than Astra and Sol for technical communication, similar on vision and computer use but much faster.
Read original source ↗ - DecryptSource report revisedOpenAI launches GPT-6.1 Terra, a cheaper AI model with improved reasoning and computer use
Report first recorded Oct 27, 06:14
OpenAI released GPT-6.1 Terra on October 26, a mid-tier model priced at 50% less than GPT-5.6 Terra. In internal testing, the model reduced its factual error rate from 53.6% to 16.7% at low reasoning effort, and at maximum effort it solved 50.3% of Terminal-Bench 4.0 coding tasks, up from 21.5% for the previous version. OpenAI also claims Terra completes tasks over three times faster than GPT-6 Astra and GPT-6.1 Sol.
Read original source ↗ - The DecoderSelectedSource report revisedOpenAI launches GPT-6.1 Terra with 50% lower API prices and improved reasoning performance
Report first recorded Oct 27, 05:26
OpenAI has released GPT-6.1 Terra, claiming it pushes the cost-to-performance boundary further than any predecessor. API prices are half those of GPT-5.6 Terra, with input tokens at $1 per million and output tokens at $5. Cached inputs drop to $0.05 per million. On the Intelligence Index, Terra scores 48, a 6-point gain over GPT-5.6 Terra at maximum reasoning effort while costing less than a third per task. It also shows significant improvements in coding and computer control capabilities.
Read original source ↗
- OpenAI NewsFirst-party sourceSource report revisedOpenAI releases GPT-6.1 Terra with enhanced computer use capabilities
Report first recorded Oct 27, 03:24
OpenAI announced GPT-6.1 Terra, a new model featuring improved computer use capabilities and faster processing speeds. The model achieves higher scores in benchmark tests and offers a longer context window, now available via API.
Read original source ↗
Event activity trend
- Latest comparable
- 3
- Comparable peak
- 19Oct 28, 14:00
- 24h change
- -51%
The chart compares consistently observed participants only. Hover or tap to inspect each hour; use arrow keys to move between observations.
Event discussion
Reader opinions are hosted on GitHub Discussions; posting requires a GitHub account. Comments are not verified facts.