LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

GPT-5.4 Mini vs GPT-5.6 Terra

Updated October 21, 2026. 84 benchmarks have a result for both. Coding is ranked on the Vibe Coding Index; every other board is shown as published.

Replace GPT-5.4 Mini Replace GPT-5.6 Terra
GPT-5.4 MiniOpenAI
Coding
28.2
Overall
24.1
Output $/M
$4.50
2
GPT-5.6 TerraOpenAI
Coding
69.0
Overall
42.1
Output $/M
$12.00
82
vs boards won of 84 shared

Decision reading

Pick GPT-5.6 Terra for coding: it scores 69.0 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini.

Pick GPT-5.4 Mini when output tokens cost 2.7x less, or when what matters is SAGE (Vals AI) or IFBench, where it leads.

Model AGPT-5.4 MiniOpenAI28.2/1001 independent benchmark$0.70 per solved taskModel BGPT-5.6 TerraOpenAI69.0/1002 independent benchmarks$0.42 per solved task
Shared results
84
GPT-5.4 Mini only
4
GPT-5.6 Terra only
40
Like-for-like categories
7 of 9

Shared results can support a head-to-head reading. Results only one model has describe coverage, not superiority.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
GPT-5.4 Mini 28.2 from 1 benchmark, GPT-5.6 Terra 69.0 from 2 benchmarks.
Widest category
Math, 29.3 points apart on average over 6 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5.4 Mini 18.9% against GPT-5.6 Terra 83.9%, 65.0 points apart on a 0 to 100 scale.
Output price
GPT-5.4 Mini $4.50 against GPT-5.6 Terra $12.00 per million output tokens, 2.7x apart.
Context window
GPT-5.4 Mini takes 400K tokens, GPT-5.6 Terra 1.05M.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryGPT-5.4 MiniBarsGPT-5.6 TerraTies
Coding018
Intelligence02
Agentic work06
Reasoning014
Math06
Knowledge09
General227
All shared boards282
CodingLike-for-likeFour of these boards feed the Vibe Coding Index018
Coding benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
LiveBench codingNot measured86.2%Only GPT-5.6 Terra measured
LMArena WebDev13971522GPT-5.6 Terra
AA Coding Index56.176.7GPT-5.6 Terra
LiveCodeBench (Vals AI)81.5%85.9%GPT-5.6 Terra
SWE-bench Verified (Vals AI)73.0%95.4%GPT-5.6 Terra
Terminal-Bench 4 (Vals AI)2.5%22.7%GPT-5.6 Terra
Vibe Code Bench (Vals AI)48.0%74.6%GPT-5.6 Terra
Code migration (Vals AI)12.9%47.8%GPT-5.6 Terra
ProgramBench (Vals AI)0.0%0.5%GPT-5.6 Terra
IOI (Vals AI)Not measured87.6%Only GPT-5.6 Terra measured
Terminal-Bench 2.1 (Vals AI)54.7%77.5%GPT-5.6 Terra
Vibe Code Bench 1-100 (Vals AI)Not measured14.8%Only GPT-5.6 Terra measured
SkillsBench (Vals AI)Not measured58.9%Only GPT-5.6 Terra measured
CyberBench (Vals AI)Not measured72.1%Only GPT-5.6 Terra measured
SciCode52.1%55.0%GPT-5.6 Terra
Terminal-Bench Hard (Artificial Analysis run)52.3%62.9%GPT-5.6 Terra
FrontierCode27.0%41.3%GPT-5.6 Terra
WeirdML60.3%78.3%GPT-5.6 Terra
ALE-Bench1188.61951.4GPT-5.6 Terra
LMArena Text: coding14381485GPT-5.6 Terra
LMArena Text: software and IT14361482GPT-5.6 Terra
LMArena WebDev: React13861513GPT-5.6 Terra
LMArena WebDev: HTML14191526GPT-5.6 Terra
LMArena WebDev: from an imageNot measured1541Only GPT-5.6 Terra measured
Neither model has a result on the other 14 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
AA Intelligence Index24.142.1GPT-5.6 Terra
Epoch Capabilities Index148.7160.1GPT-5.6 Terra
Agentic workLike-for-likePublic boards, shown as published, kept out of the index06
Agentic work benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
AA Agentic Index17.943.2GPT-5.6 Terra
Design Arena: agents building web appsNot measured1120Only GPT-5.6 Terra measured
Design Arena: agents building full-stack appsNot measured1061Only GPT-5.6 Terra measured
Design Arena: agents building mobile appsNot measured1112Only GPT-5.6 Terra measured
Design Arena: agents building Android appsNot measured1066Only GPT-5.6 Terra measured
Design Arena: agents building gamesNot measured1082Only GPT-5.6 Terra measured
Design Arena: HTML slidesNot measured1108Only GPT-5.6 Terra measured
Design Arena: PowerPoint slidesNot measured1130Only GPT-5.6 Terra measured
LiveBench researchNot measured60.4%Only GPT-5.6 Terra measured
LiveBench browserNot measured84.0%Only GPT-5.6 Terra measured
LiveBench long agenticNot measured48.4%Only GPT-5.6 Terra measured
SREBench (Vals AI)Not measured9.5%Only GPT-5.6 Terra measured
Finance Agent (Vals AI)45.4%54.4%GPT-5.6 Terra
Tax Agent Bench (Vals AI)8.8%30.0%GPT-5.6 Terra
Legal Research Bench (Vals AI)12.5%41.4%GPT-5.6 Terra
Excel Modeling Benchmark (Vals AI)45.4%66.2%GPT-5.6 Terra
Terminal-Bench-Science (Vals AI)Not measured10.0%Only GPT-5.6 Terra measured
τ²-Bench Telecom83.3%86.3%GPT-5.6 Terra
APEX-AgentsNot measured58.2%Only GPT-5.6 Terra measured
BALROGNot measured53.2%Only GPT-5.6 Terra measured
DeepResearch Bench36.3%Not measuredOnly GPT-5.4 Mini measured
Vending-Bench 2Not measured$7,343Only GPT-5.6 Terra measured
Neither model has a result on the other 17 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index014
Reasoning benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
GPQA Diamond83.1%90.9%GPT-5.6 Terra
LiveBench reasoningNot measured90.6%Only GPT-5.6 Terra measured
MMMU Pro (Vals AI)79.3%86.5%GPT-5.6 Terra
GPQA Diamond (Artificial Analysis run)87.5%92.5%GPT-5.6 Terra
Humanity's Last Exam (Artificial Analysis run)28.1%42.9%GPT-5.6 Terra
GPQA Diamond (Epoch AI run)86.9%93.3%GPT-5.6 Terra
ARC-AGI63.7%96.5%GPT-5.6 Terra
ARC-AGI-218.9%83.9%GPT-5.6 Terra
SimpleBenchNot measured48.9%Only GPT-5.6 Terra measured
Chess puzzles23.0%47.0%GPT-5.6 Terra
Mystery game puzzles11.0%35.0%GPT-5.6 Terra
LMArena Text: hard prompts14241469GPT-5.6 Terra
LMArena Text: expert14351496GPT-5.6 Terra
LMArena Text: hard prompts in English14201465GPT-5.6 Terra
LMArena Vision: diagrams12661295GPT-5.6 Terra
LMArena Vision: homework12851312GPT-5.6 Terra
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index06
Math benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
AIME95.6%Not measuredOnly GPT-5.4 Mini measured
OTIS Mock AIME88.9%99.7%GPT-5.6 Terra
FrontierMath, tiers 1 to 351.2%86.0%GPT-5.6 Terra
FrontierMath, tier 49.8%70.7%GPT-5.6 Terra
ProofBench21.0%74.0%GPT-5.6 Terra
LMArena Text: math14181465GPT-5.6 Terra
LMArena Text: mathematical work14291478GPT-5.6 Terra
Neither model has a result on the other 6 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index09
Knowledge benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
MMLU Pro84.5%86.7%GPT-5.6 Terra
LegalBench (Vals AI)Not measured85.1%Only GPT-5.6 Terra measured
CaseLaw (Vals AI)51.7%Not measuredOnly GPT-5.4 Mini measured
CorpFin (Vals AI)60.9%65.3%GPT-5.6 Terra
TaxEval (Vals AI)71.2%76.2%GPT-5.6 Terra
MortgageTax (Vals AI)63.5%67.3%GPT-5.6 Terra
MedCode (Vals AI)Not measured43.4%Only GPT-5.6 Terra measured
SimpleQA Verified29.4%43.2%GPT-5.6 Terra
LMArena Text: science14171452GPT-5.6 Terra
LMArena Text: medicine and healthcare13951430GPT-5.6 Terra
LMArena Text: legal and government14141461GPT-5.6 Terra
LMArena Text: business and finance14181448GPT-5.6 Terra
Neither model has a result on the other 5 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
Web design benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
Design Arena: websitesNot measured1269Only GPT-5.6 Terra measured
Design Arena: UI componentsNot measured1259Only GPT-5.6 Terra measured
Design Arena: game devNot measured1230Only GPT-5.6 Terra measured
Design Arena: data vizNot measured1240Only GPT-5.6 Terra measured
Design Arena: 3DNot measured1135Only GPT-5.6 Terra measured
Design Arena: all code categoriesNot measured1241Only GPT-5.6 Terra measured
Design Arena: SVGNot measured1220Only GPT-5.6 Terra measured
Design Arena: ASCII artNot measured1177Only GPT-5.6 Terra measured
LiveBench designNot measured76.0%Only GPT-5.6 Terra measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index227
General benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
LiveBench data analysisNot measured79.3%Only GPT-5.6 Terra measured
Public Benefits Bench (Vals AI)Not measured62.4%Only GPT-5.6 Terra measured
SAGE (Vals AI)50.8%47.0%GPT-5.4 Mini
MedScribe (Vals AI)Not measured82.9%Only GPT-5.6 Terra measured
Vals Multimodal Index54.2%65.1%GPT-5.6 Terra
IFBench73.3%71.2%GPT-5.4 Mini
Long context reasoning (AA-LCR)77.0%83.0%GPT-5.6 Terra
DTBench80.0%93.3%GPT-5.6 Terra
LMCA40.8%54.9%GPT-5.6 Terra
Surface Evolver BenchNot measured83.8%Only GPT-5.6 Terra measured
Furniture assemblyNot measured54.2%Only GPT-5.6 Terra measured
LMArena Text14121448GPT-5.6 Terra
LMArena Text: instruction following14051456GPT-5.6 Terra
LMArena Text: creative writing13701413GPT-5.6 Terra
LMArena Vision12451271GPT-5.6 Terra
LMArena Text: English14121449GPT-5.6 Terra
LMArena Text: non-English14051441GPT-5.6 Terra
LMArena Text: Chinese14461514GPT-5.6 Terra
LMArena Text: German14081460GPT-5.6 Terra
LMArena Text: Japanese13751454GPT-5.6 Terra
LMArena Text: Korean13701427GPT-5.6 Terra
LMArena Text: French14411466GPT-5.6 Terra
LMArena Text: Spanish14051451GPT-5.6 Terra
LMArena Text: Russian14171452GPT-5.6 Terra
LMArena Text: Polish14151446GPT-5.6 Terra
LMArena Text: multi-turn14291453GPT-5.6 Terra
LMArena Text: longer queries14081454GPT-5.6 Terra
LMArena Text: writing and language13931431GPT-5.6 Terra
LMArena Text: entertainment and media13731418GPT-5.6 Terra
LMArena Vision: English12441272GPT-5.6 Terra
LMArena Vision: Chinese12861337GPT-5.6 Terra
LMArena Vision: reading text12531278GPT-5.6 Terra
LMArena Vision: humor12351236GPT-5.6 Terra
LMArena Vision: recognizing things1218Not measuredOnly GPT-5.4 Mini measured
LMArena Vision: creative writing12221248GPT-5.6 Terra
LMArena DocumentNot measured1472Only GPT-5.6 Terra measured
Neither model has a result on the other 10 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.4 Mini and GPT-5.6 Terra
BenchmarkGPT-5.4 MiniGPT-5.6 TerraLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

GPT-5.4 Mini$0.0030Fits in one request
GPT-5.6 Terra$0.0080Fits in one request

GPT-5.4 Mini costs less

Repository review

50K input + 3K output tokens

GPT-5.4 Mini$0.051Fits in one request
GPT-5.6 Terra$0.136Fits in one request

GPT-5.4 Mini costs less

Agent loop

200K input + 10K output tokens

GPT-5.4 Mini$0.195Fits in one request
GPT-5.6 Terra$0.520Fits in one request

GPT-5.4 Mini costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of GPT-5.4 Mini and GPT-5.6 Terra
SpecificationOpenAI GPT-5.4 MiniOpenAI GPT-5.6 Terra
MakerOpenAIOpenAI
WeightsNo public weights listedNo public weights listed
ReleasedMarch 17, 2026July 9, 2026
Context window400K1.05M
Input $/Mtok$0.75$2.00
Output $/Mtok$4.50$12.00
Coding index28.269.0
Independent measurements behind it12
Cost per solved task$0.70$0.42
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.4 Mini better than GPT-5.6 Terra for coding?

Pick GPT-5.6 Terra for coding: it scores 69.0 on the Vibe Coding Index against 28.2 for GPT-5.4 Mini. Pick GPT-5.4 Mini when output tokens cost 2.7x less, or when what matters is SAGE (Vals AI) or IFBench, where it leads.

Which is cheaper, GPT-5.4 Mini or GPT-5.6 Terra?

GPT-5.4 Mini costs $0.75 per million input tokens and $4.50 per million output tokens; GPT-5.6 Terra costs $2.00 and $12.00. Per solved SWE-bench Verified task, GPT-5.4 Mini costs $0.70 and GPT-5.6 Terra $0.42.

Which has the bigger context window, GPT-5.4 Mini or GPT-5.6 Terra?

GPT-5.6 Terra, with 1.05M tokens against 400K.

How many benchmarks is the GPT-5.4 Mini vs GPT-5.6 Terra comparison based on?

84 benchmarks have a result for both models as of October 21, 2026. GPT-5.4 Mini's coding index rests on 1 independent benchmark and GPT-5.6 Terra's on 2. Where only one model has a result, the page says so instead of naming a winner.

Compare next