OPENAI · MODEL RELEASE TRACKER
GPT-6.1 Terra
GPT-6.1 Terra is a mid-range OpenAI model released in October 2026, positioned for fast visual reasoning, precise computer control (agent-style automation), and everyday office tasks with a strong price-to-performance focus.CURRENT SNAPSHOT4/5 DIMENSIONS WITH DATA
The dimensions that change the decision.
API token pricingReported input tokens: $1 per million; output tokens: $5 per million. OpenAI/coverage says these API prices are half those of GPT-5.6 Terra. Cached inputs reported at $0.05 per million.
Not established from the available sources.
ModalitiesCombines text capabilities with visual reasoning/image processing and precise computer control (agent-style automation).
Burgers PDE mean relative L2 error reduction vs fixed-prompt agentRSI-Opt with GPT-6.1 reduced mean relative L2 error on the Burgers PDE by approximately 8.8-fold compared to the fixed-prompt agent baseline (reported in the supplied source).
Intelligence Index and cost-per-task comparisonAt maximum reasoning effort, GPT-6.1 Terra scores 6 points higher on the Intelligence Index than GPT-5.6 Terra while costing less than a third per task ($0.40 vs $1.40); at lower effort levels gains of 7–13 points are reported.
Coding Agent Index v2.0 scoreScores 48 on the Coding Agent Index v2.0 (reported as six points below GPT-6.1 Sol's 54 and 17 points above GPT-6 Luna's 31).
Agents' Last Exam and speed / token-efficiency claimsReported to match GPT-6 Astra on the Agents' Last Exam at one-tenth the token price and to finish more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
Intelligence Index change vs GPT-5.6 TerraAt maximum reasoning effort, reported to gain 6 points on the Intelligence Index over GPT-5.6 Terra while costing less than a third per task (example cost $0.40 vs $1.40).
Agents' Last Exam and efficiencyReported to match GPT-6 Astra's performance on the Agents' Last Exam at roughly one‑tenth the token price; OpenAI additionally says Terra finishes more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
AvailabilityAvailable now via ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users, and available through the OpenAI API.
RELEASE TIMELINE
Published, source-backed release events only.
VERIFIABLE FACTS
Every value stays attached to a source and date.
RELEASE · Reported launch / announcementDEVELOPER CLAIM
Media reporting states that OpenAI launched a GPT-6.1 Sol AI model at its DevDay (mentioned alongside the Dots AI agent).
AVAILABILITY · AvailabilityDEVELOPER CLAIM
Available now via ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users, and available through the OpenAI API.
BENCHMARKS · Burgers PDE mean relative L2 error reduction vs fixed-prompt agentINDEPENDENTLY SUPPORTED
RSI-Opt with GPT-6.1 reduced mean relative L2 error on the Burgers PDE by approximately 8.8-fold compared to the fixed-prompt agent baseline (reported in the supplied source).
RELEASE · Release dateDEVELOPER CLAIM
Announced/released by OpenAI on 2026-10-26 as 'GPT-6.1 Terra'.
PRICING · API token pricingDEVELOPER CLAIM
Reported input tokens: $1 per million; output tokens: $5 per million. OpenAI/coverage says these API prices are half those of GPT-5.6 Terra. Cached inputs reported at $0.05 per million.
MODALITIES · ModalitiesDEVELOPER CLAIM
Combines text capabilities with visual reasoning/image processing and precise computer control (agent-style automation).
CAPABILITIES · Targeted capabilities / tasksDEVELOPER CLAIM
Positioned for everyday work tasks such as drafting proposals, interpreting technical drawings, filling out forms, and automated software control.
BENCHMARKS · Intelligence Index and cost-per-task comparisonDEVELOPER CLAIM
At maximum reasoning effort, GPT-6.1 Terra scores 6 points higher on the Intelligence Index than GPT-5.6 Terra while costing less than a third per task ($0.40 vs $1.40); at lower effort levels gains of 7–13 points are reported.
BENCHMARKS · Coding Agent Index v2.0 scoreDEVELOPER CLAIM
Scores 48 on the Coding Agent Index v2.0 (reported as six points below GPT-6.1 Sol's 54 and 17 points above GPT-6 Luna's 31).
BENCHMARKS · Agents' Last Exam and speed / token-efficiency claimsDEVELOPER CLAIM
Reported to match GPT-6 Astra on the Agents' Last Exam at one-tenth the token price and to finish more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
CAPABILITIES · Target use casesDEVELOPER CLAIM
Designed for office tasks and automated software control — examples cited include drafting proposals, interpreting technical drawings, filling out forms, and automating software control; marketed as bringing fast and capable computer use to everyone.
BENCHMARKS · Intelligence Index change vs GPT-5.6 TerraDEVELOPER CLAIM
At maximum reasoning effort, reported to gain 6 points on the Intelligence Index over GPT-5.6 Terra while costing less than a third per task (example cost $0.40 vs $1.40).
BENCHMARKS · Agents' Last Exam and efficiencyDEVELOPER CLAIM
Reported to match GPT-6 Astra's performance on the Agents' Last Exam at roughly one‑tenth the token price; OpenAI additionally says Terra finishes more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
WHAT CHANGED
Stored passport versions, without reconstructed history.
BENCHMARKS · Agents' Last Exam and efficiency+ Reported to match GPT-6 Astra's performance on the Agents' Last Exam at roughly one‑tenth the token price; OpenAI additionally says Terra finishes more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
PRICING · API token pricing (examples)− OpenAI set input tokens at $1 per million and output tokens at $5 per million for GPT-6.1 Terra; cached inputs at $0.05 per million. Overall API prices are reported as half those of GPT-5.6 Terra.
AVAILABILITY · Availability+ Available now via ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users, and available through the OpenAI API.
AVAILABILITY · Availability channels− Available for Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, and accessible through the API.
BENCHMARKS · Coding Agent Index v2.0 scoreScores 48 on the Coding Agent Index v2.0 (reported six points below GPT-6.1 Sol and 17 points above GPT-6 Luna).→Scores 48 on the Coding Agent Index v2.0 (reported as six points below GPT-6.1 Sol's 54 and 17 points above GPT-6 Luna's 31).
BENCHMARKS · Intelligence Index change vs GPT-5.6 Terra+ At maximum reasoning effort, reported to gain 6 points on the Intelligence Index over GPT-5.6 Terra while costing less than a third per task (example cost $0.40 vs $1.40).
MODALITIES · Modalities+ Combines text capabilities with visual reasoning/image processing and precise computer control (agent-style automation).
CAPABILITIES · Target use cases+ Designed for office tasks and automated software control — examples cited include drafting proposals, interpreting technical drawings, filling out forms, and automating software control; marketed as bringing fast and capable computer use to everyone.
RELEASE · Release dateOpenAI released (introduced) GPT-6.1 Terra (announced 2026-10-26).→Announced/released by OpenAI on 2026-10-26 as 'GPT-6.1 Terra'.
PRICING · API token pricing+ Reported input tokens: $1 per million; output tokens: $5 per million. OpenAI/coverage says these API prices are half those of GPT-5.6 Terra. Cached inputs reported at $0.05 per million.
MODALITIES · Modalities / abilities described by developer− Described as combining fast visual reasoning with precise computer control (plus a more natural writing style).
BENCHMARKS · Agents' Last Exam and speed / token-efficiency claims+ Reported to match GPT-6 Astra on the Agents' Last Exam at one-tenth the token price and to finish more than three times faster than GPT-6 Astra and GPT-6.1 Sol.
PRICING · API token pricing (examples)+ OpenAI set input tokens at $1 per million and output tokens at $5 per million for GPT-6.1 Terra; cached inputs at $0.05 per million. Overall API prices are reported as half those of GPT-5.6 Terra.
AVAILABILITY · Availability channels+ Available for Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex, and accessible through the API.
BENCHMARKS · Coding Agent Index v2.0+ Scores 48 on the Coding Agent Index v2.0 (reported six points below GPT-6.1 Sol and 17 points above GPT-6 Luna).
BENCHMARKS · Intelligence Index and cost-per-task comparison+ At maximum reasoning effort, GPT-6.1 Terra scores 6 points higher on the Intelligence Index than GPT-5.6 Terra while costing less than a third per task ($0.40 vs $1.40); at lower effort levels gains of 7–13 points are reported.
RELEASE · Release / announcement+ OpenAI released (introduced) GPT-6.1 Terra (announced 2026-10-26).
CAPABILITIES · Targeted capabilities / tasks+ Positioned for everyday work tasks such as drafting proposals, interpreting technical drawings, filling out forms, and automated software control.
AVAILABILITY · Planned variant not released− OpenAI said it would not release a planned GPT-6.1 Astra model due to worries about safety.
MODALITIES · Modalities / abilities described by developer+ Described as combining fast visual reasoning with precise computer control (plus a more natural writing style).
BENCHMARKS · Burgers PDE mean relative L2 error reduction vs fixed-prompt agent+ RSI-Opt with GPT-6.1 reduced mean relative L2 error on the Burgers PDE by approximately 8.8-fold compared to the fixed-prompt agent baseline (reported in the supplied source).