Claude Opus 4.7 vs Kimi K2.6
Wins 5 of 8 areas
Coding · Agents · Reasoning · Images and charts · Math
Wins 1 of 8 areas
Long documents
Claude Opus 4.7 is the stronger all-rounder.Kimi K2.6 is cheaper and better at long documents.
Scores updated · 55 tests both models report · How we compare
Where each one wins
Tests won in each of the eight areas we test. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
- CodingWriting and fixing software90Claude Opus 4.79 of 9 tests
- AgentsCarrying out multi-step tasks on its own61Claude Opus 4.76 of 7 tests
- MathCompetition and research-level math50Claude Opus 4.75 of 6 tests · 1 tie
- ReasoningHard problems that need careful thinking30Claude Opus 4.73 of 4 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video20Claude Opus 4.72 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts01Kimi K2.61 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
- Following instructionsDoing exactly what it is asked11Even1 each
Long documents rests on a single test.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
Kimi K2.6 costs 84% less for the same work.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where Claude Opus 4.7 pulls ahead
- Proof problems from the US Math OlympiadUSAMO 2026+18.1points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+17.2points ahead
- Short factual questions, answered correctlySimpleQA Verified+16.8points ahead
Where Kimi K2.6 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+17.4points ahead
- Finds hard-to-locate facts by browsing the webBrowseComp+7points ahead
- Reasons across sets of long documentsAA-LCR+2.3points ahead
Every test, side by side
All 55 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.7
- Terminal-Bench 2.1Claude Opus 4.7 by 17.283.165.9+17.2
- Terminal-Bench HardClaude Opus 4.7 by 7.651.543.9+7.6
- SWE-bench VerifiedClaude Opus 4.7 by 7.487.680.2+7.4
- SWE-bench ProClaude Opus 4.7 by 5.764.358.6+5.7
- LMArena · WebDevClaude Opus 4.7 by 49 rating points15581509+49 rating
- LiveBench · Agentic CodingClaude Opus 4.7 by 3.850.746.9+3.8
- SWE-bench MultilingualClaude Opus 4.7 by 3.880.576.7+3.8
- LiveBench · CodingClaude Opus 4.7 by 3.582.178.6+3.5
- SciCodeClaude Opus 4.7 by 354.551.5+3
AgentsClaude Opus 4.7
- GDPValClaude Opus 4.7 by 15.842.827+15.8
- AA IT-Bench SREClaude Opus 4.7 by 15.546.731.2+15.5
- τ-Bench V3 · BankingClaude Opus 4.7 by 11.334.623.3+11.3
- MCP AtlasClaude Opus 4.7 by 9.277.368.1+9.2
- BrowseCompKimi K2.6 by 779.386.3+7
- AA ApexAgentsClaude Opus 4.7 by 5.433.928.5+5.4
- OSWorld-VerifiedClaude Opus 4.7 by 4.97873.1+4.9
ReasoningClaude Opus 4.7
- LiveBench · ReasoningClaude Opus 4.7 by 7.887.279.4+7.8
- Humanity's Last ExamClaude Opus 4.7 by 4.842.337.5+4.8
- CritPtClaude Opus 4.7 by 4128+4
- GPQA Diamondtie91.491.1tie
FactsEven
- SimpleQA VerifiedClaude Opus 4.7 by 16.851.734.9+16.8
- AA-Omniscience · AccuracyClaude Opus 4.7 by 16.348.932.6+16.3
- AA-Omniscience · Non-hallucinationKimi K2.6 by 1.857.759.5+1.8
- Vectara HHEM hallucination ratelower is betterKimi K2.6 by 1.21210.8+1.2
Images and chartsClaude Opus 4.7
- CharXiv (RQ)Claude Opus 4.7 by 4.39186.7+4.3
- LMArena · VisionClaude Opus 4.7 by 31 rating points13131282+31 rating
- MMMU-Protie78.879.4tie
MathClaude Opus 4.7
- USAMO 2026Claude Opus 4.7 by 18.169.351.2+18.1
- FrontierMath Tiers 1-3 (v2)Claude Opus 4.7 by 1370.257.2+13
- LiveBench · MathematicsClaude Opus 4.7 by 8.692.884.3+8.6
- FrontierMath Tier 4Claude Opus 4.7 by 6.131.725.6+6.1
- HMMT Feb. 2026Claude Opus 4.7 by 1.293.992.7+1.2
- AIME 2026tie95.896.4tie
Following instructionsEven
- IFBenchKimi K2.6 by 17.458.676+17.4
- LiveBench · Instruction FollowingClaude Opus 4.7 by 2.366.764.4+2.3
Other results19 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- BabyVisionKimi K2.6 by 46.322.268.5+46.3
- GDPval-AA (Elo)Claude Opus 4.7 by 271 rating points17531482+271 rating
- HiL-Bench (Tools-allowed)Claude Opus 4.7 by 2341.718.7+23
- AA-OmniscienceClaude Opus 4.7 by 2227.35.3+22
- AA Agentic IndexClaude Opus 4.7 by 17.439.522.1+17.4
- AA IntelligenceClaude Opus 4.7 by 13.740.727+13.7
- livebench_data_analysisClaude Opus 4.7 by 13.278.365.1+13.2
- Artificial Analysis Coding IndexClaude Opus 4.7 by 11.873.661.8+11.8
- ToolathlonClaude Opus 4.7 by 9.359.350+9.3
- frontiermath_tier_4_v1Claude Opus 4.7 by 8.322.914.6+8.3
- τ²-Bench Telecom (AA run)Kimi K2.6 by 7.388.695.9+7.3
- Finance Agent v2Claude Opus 4.7 by 6.651.544.9+6.6
- LiveBenchClaude Opus 4.7 by 4.776.972.2+4.7
- vectara_avg_summary_lengthClaude Opus 4.7 by 32.4 rating points149.1116.7+32.4 rating
- livebench_languageClaude Opus 4.7 by 2.877.975.1+2.8
- Terminal-Bench 2.0Claude Opus 4.7 by 2.769.466.7+2.7
- vectara_answer_rateKimi K2.6 by 1.79899.7+1.7
- vectara_factual_consistencyKimi K2.6 by 1.28889.2+1.2
- HLE (with tools)tie54.754tie
Questions people ask
Which is better, Claude Opus 4.7 or Kimi K2.6?
Claude Opus 4.7 wins five of the eight areas we test: coding, agents, reasoning, images and charts and math. Kimi K2.6 wins long documents, and costs 84% less. They are level on facts and following instructions.
Which is better for coding?
Claude Opus 4.7. It wins 9 of the 9 coding tests both models report; Kimi K2.6 wins none.
Which is cheaper?
Claude Opus 4.7 costs $5.00 per million input tokens and $25.00 per million output tokens; Kimi K2.6 costs $0.95 and $4.00. That makes Kimi K2.6 about 84% cheaper for the same work.
How do you compare the two?
We use the 55 benchmark tests both models have published scores on. The verdict counts the 36 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 19 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.