Claude Fable 5 vs Kimi K2.5
Wins 6 of 7 areas
Coding · Agents · Reasoning · Facts · Images and charts · Long documents
Wins 1 of 7 areas
Following instructions
Claude Fable 5 is the stronger all-rounder.Kimi K2.5 is cheaper and better at following instructions.
Scores updated · 43 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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 software70Claude Fable 57 of 7 tests
- AgentsCarrying out multi-step tasks on its own60Claude Fable 56 of 6 tests
- ReasoningHard problems that need careful thinking40Claude Fable 54 of 4 tests
- FactsGetting facts right instead of making them up30Claude Fable 53 of 3 tests
- Images and chartsUnderstanding pictures, charts and video20Claude Fable 52 of 2 tests
- Long documentsFinding answers in very long texts10Claude Fable 51 of 1 test
- Following instructionsDoing exactly what it is asked01Kimi K2.51 of 1 test
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 94% 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 Fable 5 pulls ahead
- Abstract visual puzzles that people can solveARC-AGI-2+77.4points ahead
- Command-line tasks in a real terminalTerminal-Bench 2.1+38.9points ahead
- Real work tasks from 44 professionsGDPVal+38.4points ahead
Where Kimi K2.5 pulls ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+6.7points ahead
Every test, side by side
All 43 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Fable 5
- Terminal-Bench 2.1Claude Fable 5 by 38.984.645.7+38.9
- SWE-bench ProClaude Fable 5 by 29.38050.7+29.3
- Terminal-Bench HardClaude Fable 5 by 28.162.934.8+28.1
- LMArena · WebDevClaude Fable 5 by 189 rating points16251436+189 rating
- SWE-bench VerifiedClaude Fable 5 by 18.29576.8+18.2
- SWE-bench MultilingualClaude Fable 5 by 13.686.673+13.6
- SciCodeClaude Fable 5 by 126149+12
AgentsClaude Fable 5
- GDPValClaude Fable 5 by 38.455.617.2+38.4
- AA ApexAgentsClaude Fable 5 by 31.843.311.5+31.8
- τ-Bench V3 · BankingClaude Fable 5 by 23.938.114.2+23.9
- OSWorld-VerifiedClaude Fable 5 by 21.78563.3+21.7
- MCP AtlasClaude Fable 5 by 20.984.763.8+20.9
- BrowseCompClaude Fable 5 by 12.587.474.9+12.5
ReasoningClaude Fable 5
- ARC-AGI-2Claude Fable 5 by 77.489.211.8+77.4
- CritPtClaude Fable 5 by 25.528.63.1+25.5
- Humanity's Last ExamClaude Fable 5 by 24.855.530.7+24.8
- GPQA DiamondClaude Fable 5 by 4.792.687.9+4.7
FactsClaude Fable 5
- SimpleQA VerifiedClaude Fable 5 by 36.470.734.3+36.4
- AA-Omniscience · AccuracyClaude Fable 5 by 30.265.335.2+30.2
- AA-Omniscience · Non-hallucinationClaude Fable 5 by 2.136.434.3+2.1
Images and chartsClaude Fable 5
- MMMU-ProClaude Fable 5 by 8.884.275.4+8.8
- LMArena · VisionClaude Fable 5 by 56 rating points13251269+56 rating
Following instructionsKimi K2.5
- IFBenchKimi K2.5 by 6.763.570.2+6.7
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.
- GDPval-AA v2Claude Fable 5 by 714 rating points17231009+714 rating
- AA-OmniscienceClaude Fable 5 by 50.643.3-7.3+50.6
- ToolathlonClaude Fable 5 by 33.961.727.8+33.9
- ARC-AGI-1Claude Fable 5 by 33.298.565.3+33.2
- Artificial Analysis Coding IndexClaude Fable 5 by 29.776.546.8+29.7
- AA Agentic IndexClaude Fable 5 by 29.35121.7+29.3
- SWEBench Pro PublicClaude Fable 5 by 29.38050.7+29.3
- AA IntelligenceClaude Fable 5 by 26.149.623.5+26.1
- PaperBenchClaude Fable 5 by 25.388.863.5+25.3
- DeepSearchQA (F1)Claude Fable 5 by 17.194.277.1+17.1
- HLE (with tools)Claude Fable 5 by 13.663.850.2+13.6
- BrowseComp (w/ Ctx)Claude Fable 5 by 13.18874.9+13.1
- τ³-Bench BankingClaude Fable 5 by 12.626.814.2+12.6
- charxiv_rq_with_pythonClaude Fable 5 by 10.789.478.7+10.7
- Global-MMLU-LiteClaude Fable 5 by 9.393.384+9.3
- LiveBenchClaude Fable 5 by 9.278.369.1+9.2
- τ²-Bench Telecom (AA run)Claude Fable 5 by 2.698.595.9+2.6
- WideSearchClaude Fable 5 by 2.281.279+2.2
- StrongREJECTtie98.799.5tie
Questions people ask
Which is better, Claude Fable 5 or Kimi K2.5?
Claude Fable 5 wins six of the seven areas where both have results: coding, agents, reasoning, facts, images and charts and long documents. Kimi K2.5 wins following instructions, and costs 94% less.
Which is better for coding?
Claude Fable 5. It wins 7 of the 7 coding tests both models report; Kimi K2.5 wins none.
Which is cheaper?
Claude Fable 5 costs $10.00 per million input tokens and $50.00 per million output tokens; Kimi K2.5 costs $0.60 and $3.00. That makes Kimi K2.5 about 94% cheaper for the same work.
How do you compare the two?
We use the 43 benchmark tests both models have published scores on. The verdict counts the 24 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.