Claude Opus 4.8 vs Kimi K2.5
Wins 6 of 8 areas
Coding · Agents · Reasoning · Facts · Images and charts · Math
Wins 1 of 8 areas
Following instructions
Claude Opus 4.8 is the stronger all-rounder.Kimi K2.5 is cheaper and better at following instructions.
Scores updated · 48 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 software70Claude Opus 4.87 of 7 tests
- AgentsCarrying out multi-step tasks on its own60Claude Opus 4.86 of 6 tests
- ReasoningHard problems that need careful thinking40Claude Opus 4.84 of 4 tests
- FactsGetting facts right instead of making them up30Claude Opus 4.83 of 3 tests
- MathCompetition and research-level math30Claude Opus 4.83 of 3 tests
- Images and chartsUnderstanding pictures, charts and video20Claude Opus 4.82 of 2 tests
- Following instructionsDoing exactly what it is asked01Kimi K2.51 of 1 test
- Long documentsFinding answers in very long texts00Even0 each · 1 tie
Long documents and following instructions rest on a single test each.
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 88% 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 Opus 4.8 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+60.3points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+38.9points ahead
- Real work tasks from 44 professionsGDPVal+30.6points ahead
Where Kimi K2.5 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+8points ahead
Every test, side by side
All 48 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Opus 4.8
- Terminal-Bench 2.1Claude Opus 4.8 by 38.984.645.7+38.9
- Terminal-Bench HardClaude Opus 4.8 by 23.558.334.8+23.5
- SWE-bench ProClaude Opus 4.8 by 18.569.250.7+18.5
- LMArena · WebDevClaude Opus 4.8 by 120 rating points15561436+120 rating
- SWE-bench VerifiedClaude Opus 4.8 by 11.888.676.8+11.8
- SWE-bench MultilingualClaude Opus 4.8 by 11.484.473+11.4
- SciCodeClaude Opus 4.8 by 5.454.449+5.4
AgentsClaude Opus 4.8
- GDPValClaude Opus 4.8 by 30.647.817.2+30.6
- AA ApexAgentsClaude Opus 4.8 by 27.939.411.5+27.9
- OSWorld-VerifiedClaude Opus 4.8 by 20.183.463.3+20.1
- τ-Bench V3 · BankingClaude Opus 4.8 by 2034.214.2+20
- MCP AtlasClaude Opus 4.8 by 18.482.263.8+18.4
- BrowseCompClaude Opus 4.8 by 9.484.374.9+9.4
ReasoningClaude Opus 4.8
- ARC-AGI-2Claude Opus 4.8 by 60.372.111.8+60.3
- Humanity's Last ExamClaude Opus 4.8 by 1848.730.7+18
- CritPtClaude Opus 4.8 by 17.820.93.1+17.8
- GPQA DiamondClaude Opus 4.8 by 4.19287.9+4.1
FactsClaude Opus 4.8
- AA-Omniscience · Non-hallucinationClaude Opus 4.8 by 26.460.734.3+26.4
- SimpleQA VerifiedClaude Opus 4.8 by 18.75334.3+18.7
- AA-Omniscience · AccuracyClaude Opus 4.8 by 13.648.835.2+13.6
Images and chartsClaude Opus 4.8
- CharXiv (RQ)Claude Opus 4.8 by 12.489.977.5+12.4
- LMArena · VisionClaude Opus 4.8 by 26 rating points12951269+26 rating
MathClaude Opus 4.8
- HMMT Feb. 2026Claude Opus 4.8 by 8.495.587.1+8.4
- AIME 2026Claude Opus 4.8 by 4.210095.8+4.2
- IMOAnswerBenchClaude Opus 4.8 by 1.783.581.8+1.7
Other results21 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 Opus 4.8 by 584 rating points15931009+584 rating
- NL2RepoClaude Opus 4.8 by 37.769.732+37.7
- CyberGymClaude Opus 4.8 by 37.578.841.3+37.5
- AA-OmniscienceClaude Opus 4.8 by 36.128.8-7.3+36.1
- Tool-DecathlonClaude Opus 4.8 by 32.159.927.8+32.1
- ToolathlonClaude Opus 4.8 by 32.159.927.8+32.1
- Artificial Analysis Coding IndexClaude Opus 4.8 by 27.574.346.8+27.5
- ARC-AGI-1Claude Opus 4.8 by 27.292.565.3+27.2
- frontiermath_tier_4_v1Claude Opus 4.8 by 27.131.34.2+27.1
- Terminal-Bench 2.0Claude Opus 4.8 by 23.874.650.8+23.8
- AA Agentic IndexClaude Opus 4.8 by 20.942.621.7+20.9
- AA IntelligenceClaude Opus 4.8 by 18.341.823.5+18.3
- PaperBenchClaude Opus 4.8 by 16.880.363.5+16.8
- DeepSearchQA (F1)Claude Opus 4.8 by 1693.177.1+16
- ResearchRubricsClaude Opus 4.8 by 1473.559.5+14
- LiveBenchClaude Opus 4.8 by 8.177.269.1+8.1
- HLE (with tools)Claude Opus 4.8 by 7.757.950.2+7.7
- HMMT Nov. 2025Claude Opus 4.8 by 7.396.589.2+7.3
- WideSearchKimi K2.5 by 6.172.979+6.1
- τ²-Bench Telecom (AA run)Kimi K2.5 by 1.594.495.9+1.5
- matharena_visual_math_overallClaude Opus 4.8 by 181.680.6+1
Questions people ask
Which is better, Claude Opus 4.8 or Kimi K2.5?
Claude Opus 4.8 wins six of the eight areas we test: coding, agents, reasoning, facts, images and charts and math. Kimi K2.5 wins following instructions, and costs 88% less. They are level on long documents.
Which is better for coding?
Claude Opus 4.8. It wins 7 of the 7 coding tests both models report; Kimi K2.5 wins none.
Which is cheaper?
Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens; Kimi K2.5 costs $0.60 and $3.00. That makes Kimi K2.5 about 88% cheaper for the same work.
How do you compare the two?
We use the 48 benchmark tests both models have published scores on. The verdict counts the 27 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 21 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.