Claude Sonnet 5 vs Kimi K2.5
Wins 6 of 6 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents
Wins 0 of 6 areas
—
Claude Sonnet 5 is the stronger all-rounder.Kimi K2.5 is cheaper.
Scores updated · 29 tests both models report · How we compare
Where each one wins
Tests won in each of the six areas where both have results. 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 software60Claude Sonnet 56 of 6 tests
- AgentsCarrying out multi-step tasks on its own40Claude Sonnet 54 of 4 tests
- ReasoningHard problems that need careful thinking30Claude Sonnet 53 of 3 tests
- FactsGetting facts right instead of making them up20Claude Sonnet 52 of 3 tests · 1 tie
- Images and chartsUnderstanding pictures, charts and video20Claude Sonnet 52 of 3 tests · 1 tie
- Long documentsFinding answers in very long texts10Claude Sonnet 51 of 1 test
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 K2.5 costs 70% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Sonnet 5 pulls ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+34.8points ahead
- Real work tasks from 44 professionsGDPVal+31.1points ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+26.3points ahead
Where Kimi K2.5 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 29 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 5
- Terminal-Bench 2.1Claude Sonnet 5 by 34.880.545.7+34.8
- SWE-bench ProClaude Sonnet 5 by 12.563.250.7+12.5
- LMArena · WebDevClaude Sonnet 5 by 105 rating points15411436+105 rating
- SWE-bench VerifiedClaude Sonnet 5 by 8.485.276.8+8.4
- SciCodeClaude Sonnet 5 by 5.354.349+5.3
- SWE-bench MultilingualClaude Sonnet 5 by 5.378.373+5.3
AgentsClaude Sonnet 5
- GDPValClaude Sonnet 5 by 31.148.317.2+31.1
- τ-Bench V3 · BankingClaude Sonnet 5 by 23.137.314.2+23.1
- OSWorld-VerifiedClaude Sonnet 5 by 17.981.263.3+17.9
- BrowseCompClaude Sonnet 5 by 9.884.774.9+9.8
ReasoningClaude Sonnet 5
- CritPtClaude Sonnet 5 by 13.816.93.1+13.8
- Humanity's Last ExamClaude Sonnet 5 by 10.641.330.7+10.6
- GPQA DiamondClaude Sonnet 5 by 3.291.187.9+3.2
FactsClaude Sonnet 5
- AA-Omniscience · Non-hallucinationClaude Sonnet 5 by 26.360.634.3+26.3
- AA-Omniscience · AccuracyClaude Sonnet 5 by 4.94035.2+4.9
- SimpleQA Verifiedtie33.734.3tie
Images and chartsClaude Sonnet 5
- CharXiv (RQ)Claude Sonnet 5 by 10.888.377.5+10.8
- MMMU-ProClaude Sonnet 5 by 1.977.375.4+1.9
- LMArena · Visiontie12751269tie
Other results9 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.
- GDPval-AA v2Claude Sonnet 5 by 609 rating points16181009+609 rating
- Terminal-Bench 2.0Claude Sonnet 5 by 29.680.450.8+29.6
- ToolathlonClaude Sonnet 5 by 26.554.327.8+26.5
- Artificial Analysis Coding IndexClaude Sonnet 5 by 24.771.546.8+24.7
- AA-OmniscienceClaude Sonnet 5 by 23.816.4-7.3+23.8
- AA Agentic IndexClaude Sonnet 5 by 22.644.321.7+22.6
- AA IntelligenceClaude Sonnet 5 by 14.738.223.5+14.7
- LVBenchKimi K2.5 by 7.468.575.9+7.4
- HLE (with tools)Claude Sonnet 5 by 7.257.450.2+7.2
Questions people ask
Which is better, Claude Sonnet 5 or Kimi K2.5?
Claude Sonnet 5 wins all six areas where both have results: coding, agents, reasoning, facts, images and charts and long documents. Kimi K2.5 wins none, but costs 70% less.
Which is better for coding?
Claude Sonnet 5. It wins 6 of the 6 coding tests both models report; Kimi K2.5 wins none.
Which is cheaper?
Claude Sonnet 5 costs $2.00 per million input tokens and $10.00 per million output tokens; Kimi K2.5 costs $0.60 and $3.00. That makes Kimi K2.5 about 70% cheaper for the same work.
How do you compare the two?
We use the 29 benchmark tests both models have published scores on. The verdict counts the 20 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 9 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.