Enter opens the highlighted result. Esc closes.

GPT-6 Astra vs Kimi K3

GPT-6 Astra leads Kimi K3 on coding, 85.1 to 62.3 on the Vibe Coding Index, and Kimi K3 costs 3.3x less per output token ($15.00 vs $50.00 per million). Updated October 22, 2026; 100 benchmarks have a result for both.

Replace GPT-6 Astra Replace Kimi K3
GPT-6 AstraOpenAI
Coding
85.1
Overall
83.0
Output $/M
$50.00
60
Kimi K3Moonshot AI
Coding
62.3
Overall
66.5
Output $/M
$15.00
40
vs boards won of 100 shared

GPT-6 Astra or Kimi K3: which to pick

Pick GPT-6 Astra for coding: it scores 85.1 on the Vibe Coding Index against 62.3 for Kimi K3, though on 5 benchmarks against 6.

Pick Kimi K3 for output tokens 3.3x cheaper ($15.00 vs $50.00 per million) and work measured by SWE-bench Multimodal (Vals AI) or CyberBench (Vals AI), where it leads.

Model AGPT-6 AstraOpenAI85.1/1005 independent benchmarksNo cost per solved taskModel BKimi K3Moonshot AI62.3/1006 independent benchmarks$0.81 per solved task
Shared results
100
GPT-6 Astra only
21
Kimi K3 only
17
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-6 Astra 85.1 from 5 benchmarks, Kimi K3 62.3 from 6 benchmarks.
Widest category
Agentic work, 22.6 points apart on average over 20 shared benchmarks.
Widest single gap
Time Horizon Index (Vals AI): GPT-6 Astra 90.5% against Kimi K3 10.5%, 80.0 points apart on a 0 to 100 scale.
Output price
GPT-6 Astra $50.00 against Kimi K3 $15.00 per million output tokens, 3.3x apart.
Context window
GPT-6 Astra takes 1.05M tokens, Kimi K3 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-6 AstraBarsKimi K3Ties
Coding175
Intelligence20
Agentic work164
Reasoning83
Math42
Knowledge15
Web design71
General520
All shared boards6040
CodingLike-for-likeSix of these boards feed the Vibe Coding Index175
Coding benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
LiveBench coding93.7%59.3%GPT-6 Astra
LMArena WebDev17741655GPT-6 Astra
AA Coding Index76.976.2GPT-6 Astra
LiveCodeBench (Vals AI)Not measured87.2%Only Kimi K3 measured
SWE-bench Verified (Vals AI)Not measured93.4%Only Kimi K3 measured
Terminal-Bench 4 (Vals AI)59.6%17.2%GPT-6 Astra
Vibe Code Bench (Vals AI)89.6%85.0%GPT-6 Astra
Code migration (Vals AI)67.7%16.1%GPT-6 Astra
ProgramBench (Vals AI)5.5%2.0%GPT-6 Astra
IOI (Vals AI)100.0%48.9%GPT-6 Astra
Terminal-Bench 2.1 (Vals AI)87.3%80.9%GPT-6 Astra
SWE-bench Multimodal (Vals AI)43.3%44.2%Kimi K3
Vibe Code Bench 1-100 (Vals AI)27.6%18.2%GPT-6 Astra
CyberBench (Vals AI)41.1%75.2%Kimi K3
SciCode56.5%59.5%Kimi K3
FrontierCode53.3%44.2%GPT-6 Astra
FrontierSWE65.5%25.9%GPT-6 Astra
GSO-Bench79.4%Not measuredOnly GPT-6 Astra measured
WeirdML93.3%82.6%GPT-6 Astra
MirrorCode46.7%Not measuredOnly GPT-6 Astra measured
ALE-Bench2951.31524.5GPT-6 Astra
LMArena Text: coding14891508Kimi K3
LMArena Text: software and IT14741499Kimi K3
LMArena WebDev: React17981670GPT-6 Astra
LMArena WebDev: HTML17751644GPT-6 Astra
LMArena WebDev: from an image17261579GPT-6 Astra
SciCode12.3%Not measuredOnly GPT-6 Astra measured
Neither model has a result on the other 12 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
Intelligence benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
AA Intelligence Index52.743.6GPT-6 Astra
Epoch Capabilities Index168.8157.9GPT-6 Astra
Agentic workLike-for-likePublic boards, shown as published, kept out of the index164
Agentic work benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
AA Agentic Index51.050.0GPT-6 Astra
Design Arena: agents building web apps13731295GPT-6 Astra
Design Arena: agents building full-stack apps13571315GPT-6 Astra
Design Arena: agents building mobile apps13121269GPT-6 Astra
Design Arena: agents building Android appsNot measured1253Only Kimi K3 measured
Design Arena: agents building Godot games13481223GPT-6 Astra
Design Arena: agents building gamesNot measured1252Only Kimi K3 measured
Design Arena: HTML slidesNot measured1206Only Kimi K3 measured
Design Arena: PowerPoint slides12191247Kimi K3
Design Arena: spreadsheets13651070GPT-6 Astra
LiveBench research85.2%79.0%GPT-6 Astra
LiveBench browser84.0%16.0%GPT-6 Astra
LiveBench long agentic82.4%42.8%GPT-6 Astra
SREBench (Vals AI)56.9%Not measuredOnly GPT-6 Astra measured
Finance Agent (Vals AI)53.5%53.1%GPT-6 Astra
Tax Agent Bench (Vals AI)20.7%33.4%Kimi K3
Legal Research Bench (Vals AI)39.4%46.1%Kimi K3
Harvey's Legal Agent Benchmark (Vals AI)7.1%12.9%Kimi K3
Excel Modeling Benchmark (Vals AI)71.7%66.7%GPT-6 Astra
BioMysteryBench (Vals AI)79.3%72.6%GPT-6 Astra
Terminal-Bench-Science (Vals AI)62.9%1.4%GPT-6 Astra
Vals Science Index65.1%Not measuredOnly GPT-6 Astra measured
Time Horizon Index (Vals AI)90.5%10.5%GPT-6 Astra
CUA-bench (Vals AI)19.2%Not measuredOnly GPT-6 Astra measured
Robo-bench (Vals AI)23.3%Not measuredOnly GPT-6 Astra measured
APEX-Agents64.7%47.8%GPT-6 Astra
Remote Labor Index20.8%Not measuredOnly GPT-6 Astra measured
BALROG68.3%Not measuredOnly GPT-6 Astra measured
PostTrainBenchNot measured32.0%Only Kimi K3 measured
SpireClimb25.7%Not measuredOnly GPT-6 Astra measured
Vending-Bench 2$15,515$5,165GPT-6 Astra
Neither model has a result on the other 10 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index83
Reasoning benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
GPQA DiamondNot measured92.9%Only Kimi K3 measured
LiveBench reasoning92.7%90.7%GPT-6 Astra
MysteryMechanism (Vals AI)53.1%Not measuredOnly GPT-6 Astra measured
MMMU Pro (Vals AI)Not measured88.2%Only Kimi K3 measured
GPQA Diamond (Artificial Analysis run)96.3%93.5%GPT-6 Astra
Humanity's Last Exam (Artificial Analysis run)54.7%46.9%GPT-6 Astra
GPQA Diamond (Epoch AI run)95.8%93.1%GPT-6 Astra
ARC-AGI98.5%94.5%GPT-6 Astra
ARC-AGI-295.0%60.4%GPT-6 Astra
SimpleBenchNot measured60.7%Only Kimi K3 measured
Chess puzzles74.0%37.0%GPT-6 Astra
Mystery game puzzles84.0%26.0%GPT-6 Astra
EBR-bench76.2%Not measuredOnly GPT-6 Astra measured
LMArena Text: hard prompts14631495Kimi K3
LMArena Text: expert14891520Kimi K3
LMArena Text: hard prompts in English14641496Kimi K3
LMArena Vision: diagrams1297Not measuredOnly GPT-6 Astra measured
LMArena Vision: homework1337Not measuredOnly GPT-6 Astra measured
Neither model has a result on the other 4 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index42
Math benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
OTIS Mock AIME100.0%97.2%GPT-6 Astra
FrontierMath, tiers 1 to 393.7%72.2%GPT-6 Astra
FrontierMath, tier 497.6%39.0%GPT-6 Astra
ProofBench99.0%87.0%GPT-6 Astra
LMArena Text: math14651494Kimi K3
LMArena Text: mathematical work14841505Kimi K3
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index15
Knowledge benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
MMLU ProNot measured88.0%Only Kimi K3 measured
LegalBench (Vals AI)Not measured86.2%Only Kimi K3 measured
CorpFin (Vals AI)Not measured71.6%Only Kimi K3 measured
TaxEval (Vals AI)Not measured75.7%Only Kimi K3 measured
MortgageTax (Vals AI)Not measured66.3%Only Kimi K3 measured
MedCode (Vals AI)48.5%49.4%Kimi K3
SimpleQA Verified75.6%50.6%GPT-6 Astra
LMArena Text: science14511508Kimi K3
LMArena Text: medicine and healthcare14251489Kimi K3
LMArena Text: legal and government14621503Kimi K3
LMArena Text: business and finance14391470Kimi K3
FACTS Benchmark Suite75.7%Not measuredOnly GPT-6 Astra measured
Neither model has a result on the other 6 boards here
Web designLike-for-likePublic boards, shown as published, kept out of the index71
Web design benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
Design Arena: websites13361334GPT-6 Astra
Design Arena: UI components13931341GPT-6 Astra
Design Arena: game dev14781358GPT-6 Astra
Design Arena: data viz13351351Kimi K3
Design Arena: 3D15001387GPT-6 Astra
Design Arena: all code categories13791360GPT-6 Astra
Design Arena: SVG14311302GPT-6 Astra
Design Arena: ASCII art1353Not measuredOnly GPT-6 Astra measured
LiveBench design82.0%77.2%GPT-6 Astra
GeneralLike-for-likePublic boards, shown as published, kept out of the index520
General benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
Vals Index65.0%53.0%GPT-6 Astra
LiveBench data analysis83.0%78.7%GPT-6 Astra
Public Benefits Bench (Vals AI)Not measured68.2%Only Kimi K3 measured
SAGE (Vals AI)46.4%52.8%Kimi K3
MedScribe (Vals AI)87.9%88.0%Kimi K3
Vals Multimodal IndexNot measured73.4%Only Kimi K3 measured
Long context reasoning (AA-LCR)80.7%88.7%Kimi K3
DTBench97.3%91.2%GPT-6 Astra
LMCA64.4%52.7%GPT-6 Astra
Surface Evolver BenchNot measured95.0%Only Kimi K3 measured
Furniture assembly80.0%34.2%GPT-6 Astra
LMArena Text14411476Kimi K3
LMArena Text: instruction following14551485Kimi K3
LMArena Text: creative writing14201457Kimi K3
LMArena Vision1280Not measuredOnly GPT-6 Astra measured
LMArena Text: English14471481Kimi K3
LMArena Text: non-English14321466Kimi K3
LMArena Text: Chinese14891529Kimi K3
LMArena Text: German14591490Kimi K3
LMArena Text: Japanese14051494Kimi K3
LMArena Text: Korean14281460Kimi K3
LMArena Text: French14321492Kimi K3
LMArena Text: Spanish14221472Kimi K3
LMArena Text: Russian14381485Kimi K3
LMArena Text: Polish14221458Kimi K3
LMArena Text: multi-turn14471489Kimi K3
LMArena Text: longer queries14601494Kimi K3
LMArena Text: writing and language14371471Kimi K3
LMArena Text: entertainment and media14271453Kimi K3
LMArena Vision: English1280Not measuredOnly GPT-6 Astra measured
LMArena Vision: Chinese1361Not measuredOnly GPT-6 Astra measured
LMArena Vision: reading text1292Not measuredOnly GPT-6 Astra measured
LMArena Vision: creative writing1286Not measuredOnly GPT-6 Astra measured
Neither model has a result on the other 13 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for GPT-6 Astra and Kimi K3
BenchmarkGPT-6 AstraKimi K3Leader
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-6 Astra$0.035Fits in one request
Kimi K3$0.011Fits in one request

Kimi K3 costs less

Repository review

50K input + 3K output tokens

GPT-6 Astra$0.650Fits in one request
Kimi K3$0.195Fits in one request

Kimi K3 costs less

Agent loop

200K input + 10K output tokens

GPT-6 Astra$2.50Fits in one request
Kimi K3$0.750Fits in one request

Kimi K3 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-6 Astra and Kimi K3
SpecificationOpenAI GPT-6 AstraMoonshotAI Kimi K3
MakerOpenAIMoonshot AI
WeightsNo public weights listedOpen weights
ReleasedSeptember 4, 2026July 16, 2026
Context window1.05M1.05M
Input $/Mtok$10.00$3.00
Output $/Mtok$50.00$15.00
Coding index85.162.3
Independent measurements behind it56
Cost per solved taskNot computable$0.81
SourceOpenRouter listingOpenRouter listing

Questions

Is GPT-6 Astra better than Kimi K3 for coding?

Pick GPT-6 Astra for coding: it scores 85.1 on the Vibe Coding Index against 62.3 for Kimi K3, though on 5 benchmarks against 6. Pick Kimi K3 for output tokens 3.3x cheaper ($15.00 vs $50.00 per million) and work measured by SWE-bench Multimodal (Vals AI) or CyberBench (Vals AI), where it leads.

Which is cheaper, GPT-6 Astra or Kimi K3?

GPT-6 Astra costs $10.00 per million input tokens and $50.00 per million output tokens; Kimi K3 costs $3.00 and $15.00. Cost per solved task is known for Kimi K3 only.

Which has the bigger context window, GPT-6 Astra or Kimi K3?

GPT-6 Astra, with 1.05M tokens against 1.05M.

How many benchmarks is the GPT-6 Astra vs Kimi K3 comparison based on?

100 benchmarks have a result for both models as of October 22, 2026. GPT-6 Astra's coding index rests on 5 independent benchmarks and Kimi K3's on 6. 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-6-astra-vs-kimi-k3" title="GPT-6 Astra vs Kimi K3, 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-6-astra-vs-kimi-k3">LLMBenchmark.ai, GPT-6 Astra vs Kimi K3</a></p>