Skip to main content
BenchLM
Advertise

Model comparison

GPT-5.5 vs Kimi K3

Data verified

Head-to-head evidence from 34 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

GPT-5.5
OpenAI
73.51/100
Margin
7.4pts
winning →
Kimi K3
Moonshot AI
80.96/100
0 category wins3 category wins

Public leaderboard positions: GPT-5.5 #9 (Estimated); Kimi K3 #4 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. GPT-5.5 and Kimi K3 share 34 comparable benchmark results. 3 of 8 categories are comparable. 23 results are unique to GPT-5.5; 23 to Kimi K3.

Updated July 23, 2026
Shared results
34
GPT-5.5 only
23
Kimi K3 only
23
Comparable categories
3 / 8

Pick Kimi K3 if you want the stronger benchmark profile. GPT-5.5 only becomes the better choice if its workflow or ecosystem matters more than the raw scoreboard.

Confidence note. This is a partial-evidence comparison with 34 shared benchmark results across 5 evidence categories; 3 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.

Why this result

Kimi K3 is clearly ahead on the BenchAlign aggregate, 80.96 to 73.51. The gap is large enough that you do not need to squint at the spreadsheet to see the difference.

Kimi K3's sharpest advantage is in multimodal & grounded, where it averages 78.5 against 70.4. The single biggest benchmark swing on the page is OfficeQA Pro, 54.1% to 63.3%.

GPT-5.5 is also the more expensive model on tokens at $5.00 input / $30.00 output per 1M tokens, versus $3.00 input / $15.00 output per 1M tokens for Kimi K3. That is roughly 2.0x on output cost alone. Kimi K3 gives you the larger context window at 1.05M, compared with 1M for GPT-5.5.

Category breakdown

Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.

Category scores and score margins for GPT-5.5 and Kimi K3
CategoryGPT-5.5ΔKimi K3
MultimodalGPT-5.570.4Margin→ 8.1Kimi K378.5
AgenticGPT-5.581.6Margin→ 7.9Kimi K389.5
KnowledgeGPT-5.557.8Margin→ 3.2Kimi K361.0
CodingGPT-5.558.6MarginNo overlapKimi K3Not measured
ReasoningGPT-5.585.0MarginNo overlapKimi K3Not measured
MathGPT-5.547.6MarginNo overlapKimi K3Not measured

Decisive benchmark drivers

The largest measured benchmark gaps in this matchup, with exact reported values.

More
A · GPT-5.5B · Kimi K3

  1. OfficeQA Pro

    Multimodal
    Source ↗
    A 54.1%B 63.3%
    Winner: Kimi K3Δ 9.2
    OfficeQA Pro: GPT-5.5 scored 54.1%; Kimi K3 scored 63.3%. Kimi K3 wins this benchmark.

  2. BrowseComp

    Agentic
    Source ↗
    A 84.4%B 91.2%
    Winner: Kimi K3Δ 6.8
    BrowseComp: GPT-5.5 scored 84.4%; Kimi K3 scored 91.2%. Kimi K3 wins this benchmark.

  3. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 82%B 88.3%
    Winner: Kimi K3Δ 6.3
    Terminal-Bench 2.0: GPT-5.5 scored 82%; Kimi K3 scored 88.3%. Kimi K3 wins this benchmark.

  4. HLE

    Knowledge
    Source ↗
    A 52.2%B 56%
    Winner: Kimi K3Δ 3.8
    HLE: GPT-5.5 scored 52.2%; Kimi K3 scored 56%. Kimi K3 wins this benchmark.

  5. MMMU-Pro

    Multimodal
    Source ↗
    A 81.2%B 81.6%
    Winner: Kimi K3Δ 0.4
    MMMU-Pro: GPT-5.5 scored 81.2%; Kimi K3 scored 81.6%. Kimi K3 wins this benchmark.

Operational comparison

Runtime and commercial metrics are compared only when both models have a complete sourced value.

MetricGPT-5.5Kimi K3Comparison
Input / output priceUSD per 1M tokensGPT-5.5$5 input / $30 outputKimi K3$3 input / $15 outputKimi K3 has the lower combined listed price.
Generation speedtokens per secondGPT-5.5Not availableKimi K3Not availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenGPT-5.5Not availableKimi K3Not availableA complete latency comparison is not available.
Context windowmaximum listed tokensGPT-5.51MKimi K31.05MKimi K3 lists the larger context window.

Benchmark Deep Dive

AgenticKimi K3 wins
BenchmarkGPT-5.5Kimi K3Result
Terminal-Bench 2.0Source 82%88.3%Kimi K3 leads
CyberGymSource 81.8%—Not comparable
BrowseCompSource 84.4%91.2%Kimi K3 leads
OSWorld-VerifiedSource 78.7%—Not comparable
MCP AtlasSource 75.3%84.2%Kimi K3 leads
ToolathlonSource 55.6%—Not comparable
τ²-bench resultsSource 93.9%—Not comparable
AA Agentic IndexSource 44.9%50.1%Kimi K3 leads
APEX-Agents-AASource 37.7%41.3%Kimi K3 leads
GDPval-AASource 49.5%59.0%Kimi K3 leads
GDPval-AASource 14901679Kimi K3 leads
Gert LabsSource 72.93%—Not comparable
ResearchClawBenchSource 17.0%—Not comparable
OSWorld 2.0Source 13.0%—Not comparable
JobBenchSource 42.7%52.9%Kimi K3 leads
ExploitGymSource 13.4%—Not comparable
AA BriefcaseSource 11541543Kimi K3 leads
AA AutomationBenchSource 42.1%52.7%Kimi K3 leads
AA EnterpriseOps-GymSource 46.6%45.3%GPT-5.5 leads
AA Harvey LABSource 86.3%94.6%Kimi K3 leads
AA ITBenchSource 45.8%47.7%Kimi K3 leads
AA Tau3 BankingSource 31.3%33.4%Kimi K3 leads
terminalBenchHardSource 60.6%—Not comparable
aaTerminalBench21Source 84.3%85%Kimi K3 leads
DeepSearchQASource —95.0%Not comparable
Toolathlon-VerifiedSource —73.2%Not comparable
AutomationBenchSource —30.8%Not comparable
APEX-AgentsSource —37.6%Not comparable
SpreadsheetBench 2Source —34.8%Not comparable
DECK-BenchSource —73.5%Not comparable
Coding
BenchmarkGPT-5.5Kimi K3Result
SWE-bench ProSource 58.6%—Not comparable
Terminal-Bench 2.0Source 82.0%—Not comparable
Vibe Code BenchSource 69.85%—Not comparable
React Native EvalsSource 84.7%—Not comparable
cursorBench31Source 59.2%—Not comparable
cursorBench32Source 58.4%—Not comparable
AA Coding IndexSource 74.9%76.2%Kimi K3 leads
AA-SciCodeSource 56.1%58.7%Kimi K3 leads
FrontierCode 1.1 MainSource 43.0%—Not comparable
deepSweSource —67.5%Not comparable
FrontierSWESource —81.2%Not comparable
ProgramBenchSource —77.8%Not comparable
Kimi Code Bench v2Source —72.9%Not comparable
sweMarathonSource —42%Not comparable
PostTrain BenchSource —36.6%Not comparable
MLS-Bench LiteSource —48.3%Not comparable
Reasoning
BenchmarkGPT-5.5Kimi K3Result
MRCR v2 64K-128KSource 83.1%—Not comparable
MRCR v2 128K-256KSource 87.5%—Not comparable
ARC-AGI-2Source 85%—Not comparable
AA-LCRSource 74.3%74.7%Kimi K3 leads
CritPtSource 27.1%23.4%GPT-5.5 leads
KnowledgeKimi K3 wins
BenchmarkGPT-5.5Kimi K3Result
GPQASource 93.6%93.5%GPT-5.5 leads
GPQA-DSource 93.6%93.5%GPT-5.5 leads
HLESource 52.2%56%Kimi K3 leads
HLE w/o toolsSource 41.4%43.5%Kimi K3 leads
Artificial Analysis Intelligence IndexSource 54.8%57.1%Kimi K3 leads
AA-GPQA DiamondSource 93.5%93.5%Tie
AA-HLESource 44.3%44.3%Tie
AA-Omniscience IndexSource 20.1%18.4%GPT-5.5 leads
AA-Omniscience AccuracySource 56.9%46.0%GPT-5.5 leads
AA-Omniscience Hallucination RateSource 85.5%50.9%Kimi K3 leads
Math
BenchmarkGPT-5.5Kimi K3Result
FrontierMath (legacy)Source 51.7%—Not comparable
FrontierMath v2 (Tiers 1-3)Source 51.700%—Not comparable
FrontierMath v2 (Tier 4)Source 35.400%—Not comparable
MultimodalKimi K3 wins
BenchmarkGPT-5.5Kimi K3Result
MMMU-ProSource 81.2%81.6%Kimi K3 leads
MMMU-Pro w/ PythonSource 83.2%83.4%Kimi K3 leads
OfficeQA ProSource 54.1%63.3%Kimi K3 leads
AA-MMMU-ProSource 79.9%80.5%Kimi K3 leads
Design Arena WebsiteSource 12821386Kimi K3 leads
CharXiv w/o toolsSource —84.8%Not comparable
CharXivSource —91.3%Not comparable
MathVisionSource —94.3%Not comparable
MathVision w/ PythonSource —97.8%Not comparable
BabyVision w/ PythonSource —85.7%Not comparable
ZeroBenchSource —23.0%Not comparable
ZeroBench w/ PythonSource —41.0%Not comparable
WorldVQA ForceAnswerSource —51.0%Not comparable
OmniDocBenchSource —91.1%Not comparable
PerceptionBenchSource —58.5%Not comparable
Inst. Following
BenchmarkGPT-5.5Kimi K3Result
AA-IFBenchSource 75.9%—Not comparable
Frequently Asked Questions (4)

Which is better, GPT-5.5 or Kimi K3?

Kimi K3 is ahead on BenchLM's BenchAlign leaderboard, 80.96 to 73.51. The biggest single separator in this matchup is OfficeQA Pro, where the scores are 54.1% and 63.3%.

Which is better for knowledge tasks, GPT-5.5 or Kimi K3?

Kimi K3 has the edge for knowledge tasks in this comparison, averaging 61 versus 57.8. Inside this category, AA-Omniscience Hallucination Rate is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, GPT-5.5 or Kimi K3?

Kimi K3 has the edge for agentic tasks in this comparison, averaging 89.5 versus 81.6. Inside this category, AA Briefcase is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, GPT-5.5 or Kimi K3?

Kimi K3 has the edge for multimodal and grounded tasks in this comparison, averaging 78.5 versus 70.4. Inside this category, Design Arena Website is the benchmark that creates the most daylight between them.

Related Comparisons

Explore More

OpenAI Compare Pricing Methodology Find Your Best LLM Overall Rankings
Last updated: July 23, 2026

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.