Inference
EN
OpenAI

GPT-6.1 Terra

GPT-6.1 Terra is a reasoning model from OpenAI in the GPT-6 family. 34 benchmarks count toward its index, in 6 categories.

  • Closed weights
  • Reasoning
  • Input: file · image · text
  • Released 2026-10
RankedEnough results for an index.
IndexAll categories, out of 100.77.1±3.8Median 51.0
CoverageIndex weight that has results.90%
SpeedOutput tokens per second.46/sMedian 67/s
First tokenSeconds to the first output token.3.2 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$1.00batch $0.50Median $0.60
Output / 1MUS dollars per 1M output tokens.$5.00batch $2.50 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.—Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
openai/gpt-6.1-terra
Max outputMaximum output tokens per request.
128K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.05
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
5 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Amazon BedrockCloud$1.00$5.001.05M
DigitalOceanCloud$1.00$5.001.05M
Microsoft AzureCloud$1.00$5.001.05M
OpenAIMaker$1.00$5.001.05M
OpenRouterRouter$1.00$5.001.05M

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.11 counted78.63 / 218
CodingCode writing and repair.4 counted74.412 / 209
ReasoningLogic problems and puzzles.7 counted81.35 / 227
MultimodalTasks with images and text.3 counted73.24 / 84
KnowledgeFacts and expert knowledge.6 counted70.518 / 268
MultilingualTasks in many languages.None counted——
InstructionTasks with strict prompt rules.None counted——
MathMath problems.3 counted76.97 / 168

Results

34 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
OTIS Mock AIME 2024-2025Math100.0%1 / 200
Massive Multi-discipline Multimodal Understanding ProMultimodal85.5%1 / 53
LiveBench Browser UseAgentic100.0%1 / 49
GPQA diamond · Epoch AIKnowledge94.3% ±1.510 / 217
Agents' Last ExamAgentic60.0%1 / 21
LiveBench Data AnalysisReasoning82.3%4 / 80
Mystery Game PuzzlesReasoningAbove fitFar above this model's index.74.0% ±4.44 / 80
Artificial Analysis MMMU-ProMultimodal84.7%6 / 113
FrontierMath-Tiers-1-3-v2-PrivateMath90.9% ±1.75 / 90
Agent Arena steerabilityAgentic16.64 / 54
Critical Physics TasksReasoning28.0%16 / 214
FrontierMath-Tier-4-v2-PrivateMath92.7% ±4.16 / 72
MazeBench: visual spatial reasoning in a 3D open worldAgentic16.0%4 / 48
EBR-benchReasoning78.1% ±5.13 / 31
Artificial Analysis Long Context ReasoningReasoning82.7%23 / 217
Artificial Analysis Humanity's Last ExamKnowledge46.2%24 / 222
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal73.6%5 / 39
Furniture AssemblyReasoning79.2% ±5.15 / 37
ExtractBench: schema-guided enterprise document extraction (LlamaIndex leaderboard)Multimodal94.3%4 / 26
LiveBench DesignAgentic82.3%8 / 52
MirrorCodeCoding76.7% ±9.36 / 39
Vibe Code Bench v1.1Coding84.0% ±2.919 / 123
Artificial Analysis Omniscience AccuracyKnowledge47.6%28 / 164
Terminal-Bench 4.0Agentic49.5% ±1.811 / 61
Agent Arena task outcomeAgentic9.711 / 54
LiveBench Coding 2.0Coding90.8%11 / 52
SimpleQA VerifiedKnowledge57.3% ±1.621 / 91
Artificial Analysis SciCodeCoding55.1%28 / 114
LiveBench ReasoningReasoning88.7%20 / 80
Artificial Analysis GDP.pdfAgentic28.6%6 / 22
HealthBench ProfessionalKnowledge62.7%6 / 18
Medical Long Context Reasoning (MLCR-AA)Reasoning30.6%8 / 22
Artificial Analysis Terminal-Bench v4.0Agentic55.6%9 / 22
LiveBench Long AgenticAgentic39.6%23 / 51
LiveBench ResearchAgentic73.9%25 / 52
Artificial Analysis AutomationBenchAgentic60.2%12 / 23
Multimodal Knowledge Agent BenchmarkAgentic43.8%10 / 19
Terminal-Bench-Science 0.1Agentic58.3%6 / 11
HealthBench HardKnowledge31.4%8 / 13
OSWorld 2.1 offline subset (82 tasks)Agentic74.2%7 / 10
SpireClimb (Slay the Spire 2, Ascension 10)Other34.7% ±5.51 / 9
SpireClimbOther22.3% ±3.43 / 9
SpireClimb (Slay the Spire 1, Ascension 20 Heart)Other10.0% ±3.94 / 9
Legal Research BenchLegal26.0% ±3.062 / 90
Tax Agent BenchFinance17.1% ±2.758 / 82
HLAB-Vals-1.1Legal1.7% ±1.035 / 35
HealthBench Professional raw scoreKnowledge64.2%—
HealthBench length-adjusted scoreKnowledge55.1%—
HealthBench raw scoreKnowledge51.1%—
ExtractBench Pro: long-tail document extraction challenges (LlamaIndex)Multimodal27.0%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYLiveBenchNo licence statedVals AINo licence statedLMArenaCC BY 4.0models.devMIT

History

5 releases
Index77.6 → 77.1
29 Oct 202612 Nov 2026
Output / 1M$5.00 → $5.00
29 Oct 202612 Nov 2026

Questions

How much does GPT-6.1 Terra cost?
$1.00 per 1M input tokens, $5.00 per 1M output tokens.
How good is GPT-6.1 Terra?
Its index is 77.1 out of 100. That is place 9 of 283 ranked models.
What is GPT-6.1 Terra best at?
Agentic: 78.6 out of 100, place 3 of 218.
How fast is GPT-6.1 Terra?
46 output tokens per second, 3.2 s to the first token.
What is the context window of GPT-6.1 Terra?
1.05M tokens.
Is GPT-6.1 Terra open source?
No. Its weights are not published.
When was GPT-6.1 Terra released?
October 2026.

Badge

GPT-6.1 Terra on Inference360: place 9 of 283, index 77.1
HTML
<a href="https://inference360.ai/models/gpt-6-1-terra"><img src="https://inference360.ai/badges/gpt-6-1-terra.svg" alt="GPT-6.1 Terra on Inference360: place 9 of 283, index 77.1" height="20"></a>
Markdown
[![GPT-6.1 Terra on Inference360: place 9 of 283, index 77.1](https://inference360.ai/badges/gpt-6-1-terra.svg)](https://inference360.ai/models/gpt-6-1-terra)

Compare with

Claude Sonnet 5.577.2Muse Wonder Full77.3GPT-6.1 Sol78.0GPT-6 Astra78.0Claude Fable 5.176.2

More from OpenAI

All models
GPT-6.1 Terra ProUnrankedGPT-6 CyberUnrankedGPT-6 Luna (Decisions)UnrankedGPT-6.1 Sol ProUnrankedGPT-6.1 Sol78.0