Enter opens the highlighted result. Esc closes.

GPT-5.6 Terra vs Kimi K2.5

GPT-5.6 Terra leads Kimi K2.5 on coding, 63.9 to 38.6 on the Vibe Coding Index, and Kimi K2.5 costs 4.2x less per output token ($2.85 vs $12.00 per million). Updated October 22, 2026; 86 benchmarks have a result for both.

Replace GPT-5.6 Terra Replace Kimi K2.5
GPT-5.6 TerraOpenAI
Coding
63.9
Overall
68.1
Output $/M
$12.00
69
Kimi K2.5Moonshot AI
Coding
38.6
Overall
44.8
Output $/M
$2.85
17
vs boards won of 86 shared

GPT-5.6 Terra or Kimi K2.5: which to pick

Pick GPT-5.6 Terra for coding: it scores 63.9 on the Vibe Coding Index against 38.6 for Kimi K2.5.

Pick Kimi K2.5 for output tokens 4.2x cheaper ($2.85 vs $12.00 per million) and a lower cost per solved task ($0.24 against $0.42).

Model AGPT-5.6 TerraOpenAI63.9/1006 independent benchmarks$0.42 per solved taskModel BKimi K2.5Moonshot AI38.6/1003 independent benchmarks$0.24 per solved task
Shared results
86
GPT-5.6 Terra only
41
Kimi K2.5 only
13
Like-for-like categories
8 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

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-5.6 Terra 63.9 from 6 benchmarks, Kimi K2.5 38.6 from 3 benchmarks.
Widest category
Coding, 23.4 points apart on average over 15 shared benchmarks.
Widest single gap
ARC-AGI-2: GPT-5.6 Terra 83.9% against Kimi K2.5 11.8%, 72.1 points apart on a 0 to 100 scale.
Output price
GPT-5.6 Terra $12.00 against Kimi K2.5 $2.85 per million output tokens, 4.2x apart.
Context window
GPT-5.6 Terra takes 1.05M tokens, Kimi K2.5 262K.

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-5.6 TerraBarsKimi K2.5Ties
Coding150
Intelligence20
Agentic work45
Reasoning121
Math11
Knowledge73
Web design71
General216
All shared boards6917
CodingLike-for-likeSix of these boards feed the Vibe Coding Index150
Coding benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
SWE-bench VerifiedNot measured70.8%Only Kimi K2.5 measured
SWE-bench MultilingualNot measured67.3%Only Kimi K2.5 measured
LiveBench coding86.2%Not measuredOnly GPT-5.6 Terra measured
LMArena WebDev15221437GPT-5.6 Terra
AA Coding Index76.746.8GPT-5.6 Terra
LiveCodeBench (Vals AI)85.9%83.9%GPT-5.6 Terra
SWE-bench Verified (Vals AI)95.4%70.0%GPT-5.6 Terra
Terminal-Bench 4 (Vals AI)22.7%Not measuredOnly GPT-5.6 Terra measured
Vibe Code Bench (Vals AI)74.6%17.5%GPT-5.6 Terra
Code migration (Vals AI)47.8%7.0%GPT-5.6 Terra
ProgramBench (Vals AI)0.5%Not measuredOnly GPT-5.6 Terra measured
IOI (Vals AI)87.6%Not measuredOnly GPT-5.6 Terra measured
Terminal-Bench 2.1 (Vals AI)77.5%42.0%GPT-5.6 Terra
Vibe Code Bench 1-100 (Vals AI)14.8%Not measuredOnly GPT-5.6 Terra measured
SkillsBench (Vals AI)58.9%Not measuredOnly GPT-5.6 Terra measured
CyberBench (Vals AI)72.1%Not measuredOnly GPT-5.6 Terra measured
SciCode55.0%Not measuredOnly GPT-5.6 Terra measured
Terminal-Bench Hard (Artificial Analysis run)62.9%34.8%GPT-5.6 Terra
Terminal-BenchNot measured43.2%Only Kimi K2.5 measured
FrontierCode41.3%Not measuredOnly GPT-5.6 Terra measured
WeirdML78.3%45.6%GPT-5.6 Terra
ALE-Bench1951.4821.6GPT-5.6 Terra
LMArena Text: coding14851483GPT-5.6 Terra
LMArena Text: software and IT14821471GPT-5.6 Terra
LMArena WebDev: React15131429GPT-5.6 Terra
LMArena WebDev: HTML15261426GPT-5.6 Terra
LMArena WebDev: from an image15411439GPT-5.6 Terra
SciCode10.8%Not measuredOnly GPT-5.6 Terra measured
Neither model has a result on the other 11 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
AA Intelligence Index42.123.5GPT-5.6 Terra
Epoch Capabilities Index160.1148.1GPT-5.6 Terra
Agentic workLike-for-likePublic boards, shown as published, kept out of the index45
Agentic work benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
AA Agentic Index43.2Not measuredOnly GPT-5.6 Terra measured
Design Arena: agents building web apps11201136Kimi K2.5
Design Arena: agents building full-stack apps10611114Kimi K2.5
Design Arena: agents building mobile apps11111147Kimi K2.5
Design Arena: agents building Android apps10621118Kimi K2.5
Design Arena: agents building Godot gamesNot measured1219Only Kimi K2.5 measured
Design Arena: agents building games1083Not measuredOnly GPT-5.6 Terra measured
Design Arena: HTML slides1058Not measuredOnly GPT-5.6 Terra measured
Design Arena: PowerPoint slides1105Not measuredOnly GPT-5.6 Terra measured
LiveBench research60.4%Not measuredOnly GPT-5.6 Terra measured
LiveBench browser84.0%Not measuredOnly GPT-5.6 Terra measured
LiveBench long agentic48.4%Not measuredOnly GPT-5.6 Terra measured
SREBench (Vals AI)9.5%Not measuredOnly GPT-5.6 Terra measured
Finance Agent (Vals AI)54.4%35.8%GPT-5.6 Terra
Tax Agent Bench (Vals AI)30.0%Not measuredOnly GPT-5.6 Terra measured
Legal Research Bench (Vals AI)41.4%15.9%GPT-5.6 Terra
Excel Modeling Benchmark (Vals AI)66.2%28.5%GPT-5.6 Terra
Terminal-Bench-Science (Vals AI)10.0%Not measuredOnly GPT-5.6 Terra measured
τ²-Bench Telecom86.3%95.9%Kimi K2.5
APEX-Agents58.2%Not measuredOnly GPT-5.6 Terra measured
BALROG53.2%Not measuredOnly GPT-5.6 Terra measured
OSWorldNot measured63.3%Only Kimi K2.5 measured
Vending-Bench 2$7,343$1,198GPT-5.6 Terra
Neither model has a result on the other 18 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index121
Reasoning benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
GPQA Diamond90.9%84.1%GPT-5.6 Terra
LiveBench reasoning90.6%Not measuredOnly GPT-5.6 Terra measured
MMMU Pro (Vals AI)86.5%84.3%GPT-5.6 Terra
GPQA Diamond (Artificial Analysis run)92.5%87.9%GPT-5.6 Terra
Humanity's Last Exam (Artificial Analysis run)42.9%30.7%GPT-5.6 Terra
GPQA Diamond (Epoch AI run)93.3%Not measuredOnly GPT-5.6 Terra measured
ARC-AGI96.5%65.3%GPT-5.6 Terra
ARC-AGI-283.9%11.8%GPT-5.6 Terra
SimpleBench48.9%46.8%GPT-5.6 Terra
Chess puzzles47.0%Not measuredOnly GPT-5.6 Terra measured
Mystery game puzzles35.0%Not measuredOnly GPT-5.6 Terra measured
LMArena Text: hard prompts14691453GPT-5.6 Terra
LMArena Text: expert14961466GPT-5.6 Terra
LMArena Text: hard prompts in English14651460GPT-5.6 Terra
LMArena Vision: diagrams12951282GPT-5.6 Terra
LMArena Vision: homework13121297GPT-5.6 Terra
Kagi LLM Benchmark51.3%78.5%Kimi K2.5
Neither model has a result on the other 5 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index11
Math benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
AIMENot measured95.6%Only Kimi K2.5 measured
OTIS Mock AIME99.7%Not measuredOnly GPT-5.6 Terra measured
FrontierMath, tiers 1 to 386.0%Not measuredOnly GPT-5.6 Terra measured
FrontierMath, tier 470.7%Not measuredOnly GPT-5.6 Terra measured
ProofBench74.0%Not measuredOnly GPT-5.6 Terra measured
LMArena Text: math14651470Kimi K2.5
LMArena Text: mathematical work14781473GPT-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 index73
Knowledge benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
MMLU Pro86.7%85.9%GPT-5.6 Terra
LegalBench (Vals AI)85.1%Not measuredOnly GPT-5.6 Terra measured
CaseLaw (Vals AI)Not measured58.7%Only Kimi K2.5 measured
CorpFin (Vals AI)65.3%68.3%Kimi K2.5
TaxEval (Vals AI)76.2%74.2%GPT-5.6 Terra
MortgageTax (Vals AI)67.3%66.5%GPT-5.6 Terra
MedQA (Vals AI)Not measured94.4%Only Kimi K2.5 measured
MedCode (Vals AI)43.4%39.3%GPT-5.6 Terra
SimpleQA Verified43.2%34.3%GPT-5.6 Terra
LMArena Text: science14521459Kimi K2.5
LMArena Text: medicine and healthcare14301448Kimi K2.5
LMArena Text: legal and government14611441GPT-5.6 Terra
LMArena Text: business and finance14481427GPT-5.6 Terra
FACTS Benchmark Suite52.0%Not measuredOnly GPT-5.6 Terra measured
Neither model has a result on the other 4 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index71
Web design benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
Design Arena: websites12681248GPT-5.6 Terra
Design Arena: UI components12561234GPT-5.6 Terra
Design Arena: game dev12291206GPT-5.6 Terra
Design Arena: data viz12401219GPT-5.6 Terra
Design Arena: 3D11341204Kimi K2.5
Design Arena: all code categories12411236GPT-5.6 Terra
Design Arena: SVG12191149GPT-5.6 Terra
Design Arena: ASCII art11771166GPT-5.6 Terra
LiveBench design76.0%Not measuredOnly GPT-5.6 Terra measured
GeneralLike-for-likePublic boards, shown as published, kept out of the index216
General benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
LiveBench data analysis79.3%Not measuredOnly GPT-5.6 Terra measured
Public Benefits Bench (Vals AI)62.4%Not measuredOnly GPT-5.6 Terra measured
SAGE (Vals AI)47.0%49.9%Kimi K2.5
MedScribe (Vals AI)82.9%76.4%GPT-5.6 Terra
Vals Multimodal Index65.1%Not measuredOnly GPT-5.6 Terra measured
IFBench71.2%70.2%GPT-5.6 Terra
Long context reasoning (AA-LCR)83.0%78.0%GPT-5.6 Terra
DTBench93.3%Not measuredOnly GPT-5.6 Terra measured
LMCA54.9%Not measuredOnly GPT-5.6 Terra measured
Fiction.LiveBenchNot measured86.1%Only Kimi K2.5 measured
CL-benchNot measured19.3%Only Kimi K2.5 measured
Surface Evolver Bench83.8%Not measuredOnly GPT-5.6 Terra measured
Furniture assembly54.2%Not measuredOnly GPT-5.6 Terra measured
CL-bench LifeNot measured13.2%Only Kimi K2.5 measured
LMArena Text14481445GPT-5.6 Terra
LMArena Text: instruction following14561431GPT-5.6 Terra
LMArena Text: creative writing14131422Kimi K2.5
LMArena Vision12711269GPT-5.6 Terra
LMArena Text: English14491453Kimi K2.5
LMArena Text: non-English14411433GPT-5.6 Terra
LMArena Text: Chinese15141494GPT-5.6 Terra
LMArena Text: German14601441GPT-5.6 Terra
LMArena Text: Japanese14541423GPT-5.6 Terra
LMArena Text: Korean14271412GPT-5.6 Terra
LMArena Text: French14661454GPT-5.6 Terra
LMArena Text: Spanish14511450GPT-5.6 Terra
LMArena Text: Russian14521435GPT-5.6 Terra
LMArena Text: Polish14461445GPT-5.6 Terra
LMArena Text: multi-turn14531444GPT-5.6 Terra
LMArena Text: longer queries14541445GPT-5.6 Terra
LMArena Text: writing and language14311425GPT-5.6 Terra
LMArena Text: entertainment and media14181417GPT-5.6 Terra
LMArena Vision: English12721274Kimi K2.5
LMArena Vision: Chinese13371328GPT-5.6 Terra
LMArena Vision: reading text12781276GPT-5.6 Terra
LMArena Vision: humor12361284Kimi K2.5
LMArena Vision: captioningNot measured1257Only Kimi K2.5 measured
LMArena Vision: recognizing thingsNot measured1285Only Kimi K2.5 measured
LMArena Vision: creative writing12481269Kimi K2.5
LMArena Document1472Not measuredOnly GPT-5.6 Terra measured
Neither model has a result on the other 6 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-5.6 Terra and Kimi K2.5
BenchmarkGPT-5.6 TerraKimi K2.5Leader
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-5.6 Terra$0.0080Fits in one request
Kimi K2.5$0.0020Fits in one request

Kimi K2.5 costs less

Repository review

50K input + 3K output tokens

GPT-5.6 Terra$0.136Fits in one request
Kimi K2.5$0.035Fits in one request

Kimi K2.5 costs less

Agent loop

200K input + 10K output tokens

GPT-5.6 Terra$0.520Fits in one request
Kimi K2.5$0.135Fits in one request

Kimi K2.5 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-5.6 Terra and Kimi K2.5
SpecificationOpenAI GPT-5.6 TerraMoonshotAI Kimi K2.5
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedJuly 9, 2026January 27, 2026
Context window1.05M262K
Input $/Mtok$2.00$0.532
Output $/Mtok$12.00$2.85
Coding index63.938.6
Independent measurements behind it63
Cost per solved task$0.42$0.24
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-5.6 Terra better than Kimi K2.5 for coding?

Pick GPT-5.6 Terra for coding: it scores 63.9 on the Vibe Coding Index against 38.6 for Kimi K2.5. Pick Kimi K2.5 for output tokens 4.2x cheaper ($2.85 vs $12.00 per million) and a lower cost per solved task ($0.24 against $0.42).

Which is cheaper, GPT-5.6 Terra or Kimi K2.5?

GPT-5.6 Terra costs $2.00 per million input tokens and $12.00 per million output tokens; Kimi K2.5 costs $0.532 and $2.85. Per solved SWE-bench Verified task, GPT-5.6 Terra costs $0.42 and Kimi K2.5 $0.24.

Which has the bigger context window, GPT-5.6 Terra or Kimi K2.5?

GPT-5.6 Terra, with 1.05M tokens against 262K.

How many benchmarks is the GPT-5.6 Terra vs Kimi K2.5 comparison based on?

86 benchmarks have a result for both models as of October 22, 2026. GPT-5.6 Terra's coding index rests on 6 independent benchmarks and Kimi K2.5's on 3. Where only one model has a result, the page says so instead of naming a winner.

Compare next

Embed this comparison on your site

Free under CC BY 4.0, while the source link stays visible. Preview the frame · Terms and other embeds

<iframe src="https://llmbenchmark.ai/embed/compare/gpt-5-6-terra-vs-kimi-k2-5" title="GPT-5.6 Terra vs Kimi K2.5, LLMBenchmark.ai" width="100%" height="360" style="border:0;max-width:100%" loading="lazy"></iframe>
<p style="font:13px/1.4 sans-serif;margin:6px 0 0">Source: <a href="https://llmbenchmark.ai/compare/gpt-5-6-terra-vs-kimi-k2-5">LLMBenchmark.ai, GPT-5.6 Terra vs Kimi K2.5</a></p>