Skip to main content
BenchLM
Advertise

Model comparison

Claude Opus 4.8 vs Kimi K3

Data verified

Head-to-head evidence from 31 shared benchmark results across 5 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.

78.34/100
Margin
2.6pts
winning →
Kimi K3
Moonshot AI
80.96/100
1 category wins2 category wins

Public leaderboard positions: Claude Opus 4.8 #5 (Supported); Kimi K3 #4 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.

Evidence parity. Claude Opus 4.8 and Kimi K3 share 31 comparable benchmark results. 3 of 8 categories are comparable. 22 results are unique to Claude Opus 4.8; 26 to Kimi K3.

Updated July 23, 2026
Shared results
31
Claude Opus 4.8 only
22
Kimi K3 only
26
Comparable categories
3 / 8

Pick Kimi K3 if you want the stronger benchmark profile. Claude Opus 4.8 only becomes the better choice if knowledge is the priority.

Confidence note. This is a partial-evidence comparison with 31 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 has the cleaner BenchAlign overall profile here, landing at 80.96 versus 78.34. It is a real lead, but still close enough that category-level strengths matter more than the headline number.

Kimi K3's sharpest advantage is in agentic, where it averages 89.5 against 80.3. The single biggest benchmark swing on the page is Terminal-Bench 2.0, 74.6% to 88.3%. Claude Opus 4.8 does hit back in knowledge, so the answer changes if that is the part of the workload you care about most.

Claude Opus 4.8 is also the more expensive model on tokens at $5.00 input / $25.00 output per 1M tokens, versus $3.00 input / $15.00 output per 1M tokens for Kimi K3. Kimi K3 gives you the larger context window at 1.05M, compared with 1M for Claude Opus 4.8.

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 Claude Opus 4.8 and Kimi K3
CategoryClaude Opus 4.8ΔKimi K3
AgenticClaude Opus 4.880.3Margin→ 9.2Kimi K389.5
KnowledgeClaude Opus 4.862.7Margin← 1.7Kimi K361.0
MultimodalClaude Opus 4.877.0Margin→ 1.5Kimi K378.5
CodingClaude Opus 4.881.1MarginNo overlapKimi K3Not measured
MathClaude Opus 4.853.9MarginNo overlapKimi K3Not measured

Decisive benchmark drivers

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

More
A · Claude Opus 4.8B · Kimi K3

  1. Terminal-Bench 2.0

    Agentic
    Source ↗
    A 74.6%B 88.3%
    Winner: Kimi K3Δ 13.7
    Terminal-Bench 2.0: Claude Opus 4.8 scored 74.6%; Kimi K3 scored 88.3%. Kimi K3 wins this benchmark.

  2. BrowseComp

    Agentic
    Source ↗
    A 84.3%B 91.2%
    Winner: Kimi K3Δ 6.9
    BrowseComp: Claude Opus 4.8 scored 84.3%; Kimi K3 scored 91.2%. Kimi K3 wins this benchmark.

  3. OfficeQA Pro

    Multimodal
    Source ↗
    A 66.2%B 63.3%
    Winner: Claude Opus 4.8Δ 2.9
    OfficeQA Pro: Claude Opus 4.8 scored 66.2%; Kimi K3 scored 63.3%. Claude Opus 4.8 wins this benchmark.

  4. HLE

    Knowledge
    Source ↗
    A 57.9%B 56%
    Winner: Claude Opus 4.8Δ 1.9
    HLE: Claude Opus 4.8 scored 57.9%; Kimi K3 scored 56%. Claude Opus 4.8 wins this benchmark.

  5. CharXiv

    Multimodal
    Source ↗
    A 89.9%B 91.3%
    Winner: Kimi K3Δ 1.4
    CharXiv: Claude Opus 4.8 scored 89.9%; Kimi K3 scored 91.3%. Kimi K3 wins this benchmark.

Operational comparison

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

MetricClaude Opus 4.8Kimi K3Comparison
Input / output priceUSD per 1M tokensClaude Opus 4.8$5 input / $25 outputKimi K3$3 input / $15 outputKimi K3 has the lower combined listed price.
Generation speedtokens per secondClaude Opus 4.8Not availableKimi K3Not availableA complete speed comparison is not available.
First-answer latencyseconds to first tokenClaude Opus 4.8Not availableKimi K3Not availableA complete latency comparison is not available.
Context windowmaximum listed tokensClaude Opus 4.81MKimi K31.05MKimi K3 lists the larger context window.

Benchmark Deep Dive

AgenticKimi K3 wins
BenchmarkClaude Opus 4.8Kimi K3Result
Terminal-Bench 2.0Source 74.6%88.3%Kimi K3 leads
BrowseCompSource 84.3%91.2%Kimi K3 leads
DeepSearchQASource 93.1%95.0%Kimi K3 leads
OSWorld-VerifiedSource 83.4%—Not comparable
Finance Agent v2Source 53.9%—Not comparable
GDPval-AASource 15941679Kimi K3 leads
MCP AtlasSource 82.2%84.2%Kimi K3 leads
ToolathlonSource 59.9%—Not comparable
Gert LabsSource 72.97%—Not comparable
AA Agentic IndexSource 47.2%50.1%Kimi K3 leads
τ²-bench resultsSource 94.4%—Not comparable
GDPval-AASource 54.7%59.0%Kimi K3 leads
ResearchClawBenchSource 21.1%—Not comparable
OSWorld 2.0Source 20.6%—Not comparable
AA BriefcaseSource 13471543Kimi K3 leads
AA AutomationBenchSource 48.5%52.7%Kimi K3 leads
AA EnterpriseOps-GymSource 44.0%45.3%Kimi K3 leads
AA Harvey LABSource 91.1%94.6%Kimi K3 leads
AA Tau3 BankingSource 27.6%33.4%Kimi K3 leads
terminalBenchHardSource 58.3%—Not comparable
aaTerminalBench21Source 84.6%85%Kimi K3 leads
Toolathlon-VerifiedSource —73.2%Not comparable
AutomationBenchSource —30.8%Not comparable
JobBenchSource —52.9%Not comparable
APEX-AgentsSource —37.6%Not comparable
SpreadsheetBench 2Source —34.8%Not comparable
DECK-BenchSource —73.5%Not comparable
APEX-Agents-AASource —41.3%Not comparable
AA ITBenchSource —47.7%Not comparable
Coding
BenchmarkClaude Opus 4.8Kimi K3Result
SWE-bench VerifiedSource 88.6%—Not comparable
SWE-bench ProSource 69.2%—Not comparable
SWE MultilingualSource 84.4%—Not comparable
SWE MultimodalSource 38.4%—Not comparable
Terminal-Bench 2.0Source 74.6%—Not comparable
cursorBench31Source 58.4%—Not comparable
cursorBench32Source 62.3%—Not comparable
AA Coding IndexSource 74.3%76.2%Kimi K3 leads
AA-SciCodeSource 53.5%58.7%Kimi K3 leads
FrontierCode 1.1 MainSource 46.5%—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
BenchmarkClaude Opus 4.8Kimi K3Result
AA-LCRSource 67.7%74.7%Kimi K3 leads
CritPtSource 20.9%23.4%Kimi K3 leads
KnowledgeClaude Opus 4.8 wins
BenchmarkClaude Opus 4.8Kimi K3Result
GPQASource 93.6%93.5%Claude Opus 4.8 leads
GPQA-DSource 93.6%93.5%Claude Opus 4.8 leads
HLESource 57.9%56%Claude Opus 4.8 leads
HLE w/o toolsSource 49.8%43.5%Claude Opus 4.8 leads
Artificial Analysis Intelligence IndexSource 55.7%57.1%Kimi K3 leads
AA-GPQA DiamondSource 92.0%93.5%Kimi K3 leads
AA-HLESource 45.7%44.3%Claude Opus 4.8 leads
AA-Omniscience IndexSource 27.4%18.4%Claude Opus 4.8 leads
AA-Omniscience AccuracySource 46.6%46.0%Claude Opus 4.8 leads
AA-Omniscience Hallucination RateSource 35.9%50.9%Claude Opus 4.8 leads
Math
BenchmarkClaude Opus 4.8Kimi K3Result
USAMO 2026Source 96.7%—Not comparable
FrontierMath v2 (Tiers 1-3)Source 47.241%—Not comparable
FrontierMath v2 (Tier 4)Source 31.250%—Not comparable
Multilingual
BenchmarkClaude Opus 4.8Kimi K3Result
INCLUDESource 87.6%—Not comparable
MultimodalKimi K3 wins
BenchmarkClaude Opus 4.8Kimi K3Result
OfficeQA ProSource 66.2%63.3%Claude Opus 4.8 leads
ScreenSpot ProSource 87.9%—Not comparable
CharXivSource 89.9%91.3%Kimi K3 leads
CharXiv w/o toolsSource 80.5%84.8%Kimi K3 leads
Design Arena WebsiteSource 12701386Kimi K3 leads
MMMU-ProSource —81.6%Not comparable
MMMU-Pro w/ PythonSource —83.4%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
AA-MMMU-ProSource —80.5%Not comparable
Inst. Following
BenchmarkClaude Opus 4.8Kimi K3Result
AA-IFBenchSource 62.2%—Not comparable
Frequently Asked Questions (4)

Which is better, Claude Opus 4.8 or Kimi K3?

Kimi K3 is ahead on BenchLM's BenchAlign leaderboard, 80.96 to 78.34. The biggest single separator in this matchup is Terminal-Bench 2.0, where the scores are 74.6% and 88.3%.

Which is better for knowledge tasks, Claude Opus 4.8 or Kimi K3?

Claude Opus 4.8 has the edge for knowledge tasks in this comparison, averaging 62.7 versus 61. Inside this category, AA-Omniscience Hallucination Rate is the benchmark that creates the most daylight between them.

Which is better for agentic tasks, Claude Opus 4.8 or Kimi K3?

Kimi K3 has the edge for agentic tasks in this comparison, averaging 89.5 versus 80.3. Inside this category, AA Briefcase is the benchmark that creates the most daylight between them.

Which is better for multimodal and grounded tasks, Claude Opus 4.8 or Kimi K3?

Kimi K3 has the edge for multimodal and grounded tasks in this comparison, averaging 78.5 versus 77. Inside this category, Design Arena Website is the benchmark that creates the most daylight between them.

Related Comparisons

Explore More

Anthropic 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.