Is Claude Opus 4.8 better than Kimi K3?

Claude Opus 4.8 (max) and Kimi K3 (max) are neck and neck on quality (63.3 vs 64.8, within the ±5.8 margin). The BenchLeader Index combines every independent quality benchmark; the two are within the margin of error as of 2026-10-25, so check the category scores for your use.

Which is cheaper, Claude Opus 4.8 or Kimi K3, and will that change?

Kimi K3 is cheaper today: $6.00 against $10.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.

Is Claude Opus 4.8 better than Kimi K3 for coding?

Kimi K3 scores higher in coding (65 vs 59 on the category index, where 50 is average). On SciCode, the top shared coding benchmark, it is 53.5% for Claude Opus 4.8 against 59.5% for Kimi K3.

Is Claude Opus 4.8 better than Kimi K3 for agents and tool use?

Kimi K3 scores higher in agentic tasks (59 vs 57 on the category index, where 50 is average). On MCP Atlas, the top shared agents & tools benchmark, it is 82.2% for Claude Opus 4.8 against 82.3% for Kimi K3.

Which is better for long documents?

Kimi K3 scores higher on long-context benchmarks (70 vs 64). Both take about 1M tokens: Claude Opus 4.8 1,000,000, Kimi K3 1,048,576.

Which has lower latency?

Claude Opus 4.8 (max) answers first: 29.3 s against 52.6 s to the first answer token, as Artificial Analysis measures it. Reasoning effort drives most of this: a higher setting thinks longer before it answers.

Which is faster, Claude Opus 4.8 or Kimi K3?

Claude Opus 4.8 streams faster: 63 against 41 output tokens per second.

When was Kimi K3 released?

Kimi K3 was released on 16 Jul 2026.

Claude Opus 4.8 vs Kimi K3

Full profiles: Claude Opus 4.8 (Anthropic) and Kimi K3 (Moonshot AI): every benchmark, price and rank.

Updated 25 Oct 2026 · Data from 12 sources

As of 25 Oct 2026, Claude Opus 4.8 (max) and Kimi K3 (max) are neck and neck on quality (63.3 vs 64.8 on the BenchLeader Index, inside the ±5.8 margin of error), and Kimi K3 (max) costs 1.7× less, which makes it the better buy for most workloads. Pick Claude Opus 4.8 for maths (68 vs 63); pick Kimi K3 for long context (70 vs 64).

Claude Opus 4.8 benchmarks max Anthropic

Kimi K3 benchmarks max Moonshot AI Open in the comparison tool

Which to pick for…

The pick's numbers come first. Category scores on the BenchLeader scale (50 is the average model, 15 points one standard deviation). High confidence means a lead of half a standard deviation or more; under one point is called either.

Use case	Pick	Why	Confidence
Coding	Kimi K3	65 vs 59 category score; SciCode 59.5% vs 53.5%	Medium
Agents & tool use	Kimi K3	59 vs 57 category score; MCP Atlas 82.3% vs 82.2%	Medium
Maths	Claude Opus 4.8	68 vs 63 category score; FrontierMath Tiers 1–3 80.0% vs 72.2%	Medium
Reasoning	Claude Opus 4.8	69 vs 66 category score; GPQA Diamond 91.0% vs 93.1%	Medium
Knowledge	Claude Opus 4.8	67 vs 63 category score; SimpleQA Verified 53.0% vs 50.6%	Medium
Long documents	Kimi K3	70 vs 64 category score; AA-LCR 88.7% vs 77.7%	Medium
Chat quality (human preference)	–	Only Kimi K3 has an Arena rating (68 category score)	Not comparable
Lowest cost	Kimi K3	$6.00 vs $10.00 per 1M blended; agentic coding session $0.240 vs $0.400	–
Fastest first answer	Claude Opus 4.8	29.3 s (Claude Opus 4.8 (max)) vs 52.6 s (Kimi K3 (max)) to the first answer token	–
High volume on a budget	Neither flagship	A smaller model is usually the better fit: Cheapest good LLMs, Fastest LLMs	–

Key differences

Claude Opus 4.8 (max) and Kimi K3 (max) are neck and neck on quality (63.3 vs 64.8, within the ±5.8 margin).
Claude Opus 4.8 (max) is stronger in instruction following, knowledge, maths, reasoning.
Kimi K3 (max) is stronger in agents & tools, coding, composite, long context, human preference, multimodal.
Kimi K3 (max) is 1.7× cheaper ($6.00 vs $10.00 per 1M blended).
Claude Opus 4.8 (max) streams 1.6× faster (63 vs 41 tokens per second).

At a glance

Spec	Claude Opus 4.8	Kimi K3
Developer	Anthropic	Moonshot AI
Released	28 May 2026	16 Jul 2026
Context window (tokens)	1,000,000	1,048,576
Compared at	Claude Opus 4.8 (max)	Kimi K3 (max)
Open weights	No	Yes
Price in / out / cached input per 1M	$5.00 / $25.00 / $0.500	$3.00 / $15.00 / $0.300

Category by category

Category scores on the BenchLeader scale: 50 is the average evaluated model, 15 points is one standard deviation. A gap under one point is called level.

Category	Claude Opus 4.8 (max)	Kimi K3 (max)	Leader
Reasoning	68.8	66.4	Claude Opus 4.8
Coding	58.9	64.6	Kimi K3
Agents & tools	56.5	59.4	Kimi K3
Maths	68.1	63.4	Claude Opus 4.8
Knowledge	66.9	62.8	Claude Opus 4.8
Instruction following	62.1	–	not comparable
Human preference	–	68.1	not comparable
Multimodal	–	63.6	not comparable
Long context	63.8	69.6	Kimi K3
Composite	76.2	78.3	Kimi K3

Where sources disagree

Across the 35 benchmarks both have, Claude Opus 4.8 leads 19, Kimi K3 leads 16. Every overall index that rates both puts Kimi K3 ahead.

Overall index	Claude Opus 4.8 (max)	Kimi K3 (max)	Leader
BenchLeader Index	63.3	64.8	neck and neck
AA Intelligence Index v4.3.2	41.8	43.6	Kimi K3
Arena Text	1482†	1487	Kimi K3

Each publisher runs its own tasks, harnesses and effort settings, so their overall verdicts can differ; BenchLeader reconciles them on one scale. † measured at another effort setting of the same model (the publisher did not test the one compared here).

Price, speed and latency

List prices blended at three input tokens per output token. Per-task prices use the token counts on the methodology page; suite cost is what Artificial Analysis paid to run its whole Intelligence Index.

Measure	Claude Opus 4.8 (max)	Kimi K3 (max)
Blended price per 1M tokens	$10.00	$6.00
Input / output per 1M	$5.00 / $25.00	$3.00 / $15.00
Cached input per 1M	$0.500	$0.300
Price per task: chat reply	$0.0095	$0.0057
Price per task: summarize a 30-page report	$0.075	$0.045
Price per task: code edit	$0.068	$0.041
Price per task: agentic coding session	$0.400	$0.240
Price per task: structured extraction	$0.015	$0.0090
Cost to run a full benchmark suite	$4.08	$2.00
Output speed	63 tok/s	41 tok/s
Time to first answer	29.3 s	52.6 s
Time to a complete answer	37.2 s	65.0 s

Per token is not per task: on Artificial Analysis' full suite Claude Opus 4.8 (max) cost $4.08 and Kimi K3 (max) $2.00, because Claude Opus 4.8 (max) used about 1.2× as many tokens for the same work.

Each model is compared at its best-ranked reasoning-effort setting. Claude Opus 4.8 (max) scores 63.3 and takes 29 s to its first answer token; Claude Opus 4.8 (high) scores 60.7 in 1 s, Claude Opus 4.8 scores 60.1 in 1 s. Kimi K3 (max) scores 64.8 and takes 53 s to its first answer token; Kimi K3 scores 58.8 in 1 s, Kimi K3 (low) scores 58.2 in 57 s. Higher effort means more thinking before the answer: better scores, longer waits and more tokens billed. How reasoning effort changes scores and cost.

Every benchmark

155 benchmarks, grouped by category; each name links to the board and its leaders. † measured at another effort setting of the same model, shown for reference and not counted as a win.

Metric	Publisher	Claude Opus 4.8 (max)	Kimi K3 (max)
BenchLeader Index	BenchLeader	63.3	64.8
Agents & tools score	BenchLeader	56.5	59.4
Coding score	BenchLeader	58.9	64.6
Composite score	BenchLeader	76.2	78.3
Instruction following score	BenchLeader	62.1	–
Knowledge score	BenchLeader	66.9	62.8
Long context score	BenchLeader	63.8	69.6
Maths score	BenchLeader	68.1	63.4
Reasoning score	BenchLeader	68.8	66.4
Human preference score	BenchLeader	–	68.1
Multimodal score	BenchLeader	–	63.6
Blended price $/M	BenchLeader	$10.00	$6.00
Output speed	BenchLeader	63 tok/s	41 tok/s
Time to first answer	BenchLeader	29.3 s	52.6 s
Context window	BenchLeader	1,000,000 tokens	1,048,576 tokens
Reasoning
GPQA Diamond	Epoch AI Benchmarking Hub	91.0%	93.1%
SimpleBench	SimpleBench	64.8%†	60.7%
Arena Hard Prompts	Arena (LMArena)	1505†	1516
LiveBench 2.0 Reasoning	LiveBench	89.2%†	90.7%†
LiveBench 1.0 Reasoning	LiveBench	89.2%	90.7%†
GPQA Diamond (AA)	Artificial Analysis	92.0%	93.5%
Humanity's Last Exam (AA)	Artificial Analysis	48.7%	46.9%
GPQA Diamond (Vals)	Vals AI	92.4%†	92.9%†
EnigmaEval	Scale AI SEAL	23.5%†	–
Kagi LLM Benchmark	Kagi LLM Benchmark	88.8%†	–
ARC-AGI-1	ARC Prize	92.5%	94.5%
ARC-AGI-2	ARC Prize	72.1%†	60.4%
ARC-AGI-3	ARC Prize	1.5%†	–
CritPt	Artificial Analysis	20.9%	23.4%
Chess Puzzles	Epoch AI Benchmarking Hub	52.0%	36.0%
Mystery Game Puzzles	Epoch AI Benchmarking Hub	36.0%	26.0%
LMCA	Epoch AI Benchmarking Hub	57.5%	52.7%
DTBench	Epoch AI Benchmarking Hub	94.9%	91.2%
ForecastBench	ForecastBench	66.3%	64.6%
BTF-3	Epoch AI Benchmarking Hub	14.0%†	–
Coding
SciCode	SciCode	53.5%	59.5%
WeirdML	WeirdML	70.5%†	82.6%
FrontierCode	Cognition	46.5%†	44.2%†
GSO-Bench	GSO-Bench	47.1%†	–
Arena Coding	Arena (LMArena)	1529†	1539

Show all 155 benchmarks

Metric	Publisher	Claude Opus 4.8 (max)	Kimi K3 (max)
Coding
Arena WebDev	Arena (LMArena)	1536†	1655
LiveBench 2.0 Coding	LiveBench	77.3%†	59.3%†
LiveBench 1.0 Coding	LiveBench	81.8%	81.5%†
SciCode (AA)	Artificial Analysis	54.4%	59.5%
LiveCodeBench	Vals AI	87.8%†	87.2%
IOI	Vals AI	–	48.9%
SWE-bench (Vals)	Vals AI	88.6%†	93.4%†
Code Migration	Vals AI	47.3%†	16.1%
Code Migration v1.1	Vals AI	–	28.6%
ProgramBench	Vals AI	1.0%†	2.0%†
SWE-bench Multimodal	Vals AI	–	44.2%
Vibe Code Bench 1-100	Vals AI	–	18.2%
Vibe Code Bench v1.1	Vals AI	82.7%†	85.0%
SWE Atlas: Codebase QnA	Scale AI SEAL	57.3%†	–
SWE Atlas: Refactoring	Scale AI SEAL	46.7%†	–
SWE Atlas: Test Writing	Scale AI SEAL	49.6%†	–
Surface Evolver Bench	Epoch AI Benchmarking Hub	87.5%†	93.0%
DeepSWE v1.1	Epoch AI Benchmarking Hub	59.0%	68.5%
ALE-Bench	Epoch AI Benchmarking Hub	1563.8†	1524.5
FrontierSWE	Epoch AI Benchmarking Hub	–	25.9%
Terminal-Bench 4.0 (AA)	Artificial Analysis	21.7%	12.6%
Terminal-Bench 2.1 (AA)	Artificial Analysis	84.6%	85.0%
Startrise Frontend	Startrise	–	79.8%†
FrontierCode 1.1 (AA)	Artificial Analysis	–	43.6%
Agents & tools
Terminal-Bench	Terminal-Bench	23.6%	–
OSWorld-Verified 2.0	OSWorld	20.6%	–
Remote Labor Index	Scale AI / CAIS	8.3%†	–
APEX-Agents	Mercor	48.9%	47.8%†
Arena Agent	Arena (LMArena)	–	1.6
LiveBench Long Agentic	LiveBench	48.3%†	42.8%†
LiveBench Browser	LiveBench	96.0%†	16.0%†
LiveBench Agentic Coding	LiveBench	50.5%	62.2%†
Terminal-Bench Hard	Artificial Analysis	58.3%	–
τ²-Bench Telecom (AA)	Artificial Analysis	94.4%	–
Terminal-Bench 2.1 (Vals)	Vals AI	71.9%†	80.9%†
MCP Atlas	Scale AI SEAL	82.2%	82.3%
HiL-Bench	Scale AI SEAL	35.3%†	42.7%†
τ³-bench Banking	τ²-bench	39.7%	37.1%
τ³-Banking (AA)	Artificial Analysis	34.2%	46.0%
ITBench SRE (AA)	Artificial Analysis	–	47.7%
Analyst Agent (AA)	Artificial Analysis	45.0%	38.8%
APEX-Agents (AA)	Artificial Analysis	–	41.3%
CyberBench	Vals AI	–	75.2%
Finance Agent v2	Vals AI	53.9%†	53.1%
Finance Agent v2.1	Vals AI	–	60.0%
Harvey's Legal Agent Benchmark v1.1	Vals AI	–	12.9%
Harvey's Legal Agent Benchmark v1	Vals AI	9.6%†	12.9%
Legal Research Bench	Vals AI	43.8%†	46.1%
SkillsBench	Vals AI	59.2%†	–
Tax Agent Bench	Vals AI	30.0%†	33.4%
Terminal-Bench 2.0 (Vals)	Vals AI	70.0%†	–
Terminal-Bench 4.0 (Vals)	Vals AI	23.2%†	17.2%
Terminal-Bench Science	Vals AI	4.3%†	1.4%
Time Horizon Index: KSP	Vals AI	11.8%†	10.5%
Arena Search	Arena (LMArena)	1195†	–
DeepResearch Bench	DeepResearch Bench	50.2%†	–
PostTrainBench	Epoch AI Benchmarking Hub	32.9%	32.0%†
GBAEval	Epoch AI Benchmarking Hub	70.9%†	48.4%†
Vending-Bench 2	Andon Labs	2992.3	5165.0†
AutomationBench	Artificial Analysis	45.6%	58.3%
GDP.pdf (AA)	Artificial Analysis	22.8%	22.0%
Harvey LAB-AA v1.1	Harvey	–	5.3%
EnterpriseOps-Gym	Artificial Analysis	44.0%	45.3%
AA-Briefcase v1.1	Artificial Analysis	1319	1499
Terminal-Bench-Science 0.1 (AA)	Artificial Analysis	–	8.1%
CWE-Bench (AA)	Artificial Analysis	–	53.3%
DeepSec-Bench (AA)	Artificial Analysis	–	12.4%
CyberGym end-to-end (AA)	Artificial Analysis	–	58.0%
FACTS Search	Google	79.6%†	–
Maths
FrontierMath Tiers 1–3	Epoch AI Benchmarking Hub	80.0%	72.2%
FrontierMath Tier 4	Epoch AI Benchmarking Hub	56.1%	39.0%
OTIS Mock AIME	Epoch AI Benchmarking Hub	98.3%	97.2%
ProofBench	Vals AI	69.0%	87.0%†
LiveBench Mathematics	LiveBench	94.3%	84.4%†
AIME 2026	MathArena	100.0%	96.7%†
HMMT February 2026	MathArena	95.5%	97.0%†
MathArena Apex	MathArena	81.3%	65.6%†
Arena Maths	Arena (LMArena)	1477†	1499
Knowledge
SimpleQA Verified	Epoch AI Benchmarking Hub	53.0%	50.6%
LiveBench 2.0 Data Analysis	LiveBench	66.0%†	78.7%†
LiveBench Research	LiveBench	81.8%†	79.0%†
LiveBench 1.0 Data Analysis	LiveBench	66.0%	78.7%†
AA-Omniscience	Artificial Analysis	28.8	19.7
MMLU-Pro	Vals AI	89.6%†	88.0%
LegalBench	Vals AI	83.6%†	86.2%
CorpFin	Vals AI	66.7%†	71.6%†
TaxEval	Vals AI	75.6%†	75.7%†
BioMysteryBench	Vals AI	–	72.6%
Excel Modeling Benchmark	Vals AI	69.4%†	66.7%
Excel Modeling Benchmark v1.1	Vals AI	–	71.3%
MedCode	Vals AI	53.2%†	49.4%
MedScribe	Vals AI	85.8%†	88.0%
Public Benefits Bench v1.1	Vals AI	68.1%†	68.2%
DrugDiscoveryBench	Scale AI SEAL	46.8%	–
EBR-bench	Epoch AI Benchmarking Hub	28.6%	–
GDP.pdf	Epoch AI Benchmarking Hub	24.0%	19.0%
AA-Omniscience: accuracy	Artificial Analysis	48.8%	47.6%
AA-Omniscience: non-hallucination	Artificial Analysis	60.8%	46.8%
Instruction following
LiveBench Language	LiveBench	79.7%	85.5%†
LiveBench Instruction Following	LiveBench	72.0%	71.4%†
IFBench	Artificial Analysis	62.2%	–
Arena Instruction Following	Arena (LMArena)	1478†	1489
Human preference
Arena Text	Arena (LMArena)	1474†	1487
EQ-Bench 4	EQ-Bench	1267†	1318†
Arena Creative Writing	Arena (LMArena)	1463†	1462
Arena Multi-turn	Arena (LMArena)	1495†	1502
Arena Longer Queries	Arena (LMArena)	1499†	1503
Multimodal
Arena Vision	Arena (LMArena)	1289†	–
LiveBench Design	LiveBench	78.8%†	77.2%†
MMMU-Pro	Artificial Analysis	–	80.5%
MMMU-Pro (Vals)	Vals AI	86.6%†	88.2%†
MortgageTax	Vals AI	69.9%†	66.3%†
SAGE	Vals AI	54.8%†	52.8%
Blueprint-Bench 2	Epoch AI Benchmarking Hub	14.5%†	29.5%†
AA-3D	Artificial Analysis	–	982
Long context
AA-LCR	Artificial Analysis	77.7%	88.7%
Arena Document	Arena (LMArena)	1475†	–
MLCR	Artificial Analysis	19.4%	21.7%
Composite
Epoch Capabilities Index	Epoch AI Benchmarking Hub	159.0†	157.9†
LiveBench 2.0	LiveBench	76.8%†	63.4%†
LiveBench 1.0	LiveBench	76.2%	79.2%†
AA Intelligence Index v4.3.2	Artificial Analysis	41.8	43.6
Vals Index	Vals AI	–	53.0
Vals Finance Index	Vals AI	–	58.0%
Vals RSI Index	Vals AI	24.0%	20.9%
Vals Multimodal Index	Vals AI	70.9%†	73.4%†
Vals Router Index	Vals AI	–	53.0%†
AA Cyber Index	Artificial Analysis	–	41.3
Safety & honesty
FORTRESS	Scale AI SEAL	18.2%	26.6%
Metadata
AA Openness Index	Artificial Analysis	–	38.9

Data as of 2026-10-25. Best configuration of each model; every score links to its source on the model pages.

What changed

25 Oct 2026 Claude Opus 4.8 (max): 1 new result (AA-Briefcase v1.1).
23 Oct 2026 Claude Opus 4.8 (max): 1 new result (Vals RSI Index).
23 Oct 2026 Kimi K3 (max): 7 new results (Code Migration v1.1, Excel Modeling Benchmark v1.1, Finance Agent v2.1…).
21 Oct 2026 Claude Opus 4.8 (max): 1 score revision.
21 Oct 2026 Kimi K3 (max): 1 new result (Terminal-Bench-Science 0.1 (AA)); 1 score revision.
20 Oct 2026 Claude Opus 4.8 (max): 1 new result (ForecastBench).
20 Oct 2026 Kimi K3 (max): 3 score revisions.
19 Oct 2026 Claude Opus 4.8 (max): BenchLeader Index 62.5 → 63.4.
19 Oct 2026 Kimi K3 (max): 1 score revision.
18 Oct 2026 Claude Opus 4.8 (max): 5 new results (AutomationBench, GDP.pdf (AA), MLCR…).

Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs