Claude Fable 5.2 vs Kimi K2.5
Full profiles: Claude Fable 5.2 (Anthropic) and Kimi K2.5 (Moonshot AI): every benchmark, price and rank.
Updated · Data from 10 sources
As of 9 Nov 2026, Claude Fable 5.2 is ahead of Kimi K2.5 on quality (69.4 vs 58.8 on the BenchLeader Index), but Kimi K2.5 is 13× cheaper at $1.20 vs $16.00 per 1M tokens blended. Claude Fable 5.2 leads in every category we measure both in. Claude Fable 5.2 streams ~1.5× faster.
Key facts (as of 9 Nov 2026)
- Claude Fable 5.2 scores 69.4 on the BenchLeader Index and Kimi K2.5 scores 58.8 (BenchLeader, 9 Nov 2026).
- Claude Fable 5.2 costs $16.00 per 1M tokens blended (list price, 9 Nov 2026).
- Kimi K2.5 costs $1.20 per 1M tokens blended (list price, 9 Nov 2026).
- Claude Fable 5.2 streams ~1.5× faster than Kimi K2.5 (Artificial Analysis, 7-day median, 9 Nov 2026).
- Of the 3 benchmarks both models have, Claude Fable 5.2 leads 3 (BenchLeader, 9 Nov 2026).
Claude Fable 5.2 benchmarks max Anthropic
Kimi K2.5 benchmarks thinking Moonshot AIAdd a third modelWhich 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 | Claude Fable 5.2 | 80 vs 56 category score | High |
| Reasoning | Claude Fable 5.2 | 78 vs 67 category score; Humanity's Last Exam (AA) 74.3% vs 30.7% | High |
| Knowledge | Claude Fable 5.2 | 82 vs 60 category score | High |
| Long documents | Claude Fable 5.2 | 66 vs 64 category score; AA-LCR v1.1 83.3% vs 78.0% | Medium |
| Chat quality (human preference) | – | Only Kimi K2.5 has an Arena rating (63 category score) | Not comparable |
| Lowest cost | Kimi K2.5 | $1.20 vs $16.00 per 1M blended; agentic coding session $0.048 vs $0.640 | – |
| Fastest first answer | Claude Fable 5.2 | 3.0 s (Claude Fable 5.2 (max)) vs 34.1 s (Kimi K2.5 (thinking)) 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 Fable 5.2 (max) leads on quality: 69.4 vs 58.8.
- Claude Fable 5.2 (max) is stronger in coding, knowledge, long context, reasoning.
- Kimi K2.5 (thinking) is stronger in agents & tools, human preference, instruction following, maths, multimodal.
- Kimi K2.5 (thinking) is 13× cheaper ($1.20 vs $16.00 per 1M blended).
- Claude Fable 5.2 (max) streams 1.8× faster (72 vs 39 tokens per second).
At a glance
| Spec | Claude Fable 5.2 | Kimi K2.5 |
|---|---|---|
| Developer | Anthropic | Moonshot AI |
| Released | 9 Nov 2026 | 27 Jan 2026 |
| Context window (tokens) | 1,000,000 | 262,144 |
| Compared at | Claude Fable 5.2 (max) | Kimi K2.5 (thinking) |
| Open weights | No | Yes |
| Price in / out / cached input per 1M | $8.00 / $40.00 / $0.200 | $0.600 / $3.00 / $0.100 |
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 Fable 5.2 (max) | Kimi K2.5 (thinking) | Leader |
|---|---|---|---|
| Reasoning | 78.4 | 67.3 | Claude Fable 5.2 |
| Coding | 80.2 | 55.6 | Claude Fable 5.2 |
| Agents & tools | – | 45.1 | not comparable |
| Maths | – | 56.3 | not comparable |
| Knowledge | 81.7 | 60.1 | Claude Fable 5.2 |
| Instruction following | – | 69.1 | not comparable |
| Human preference | – | 63.3 | not comparable |
| Multimodal | – | 59.2 | not comparable |
| Long context | 66.4 | 63.6 | Claude Fable 5.2 |
On one chart
Each domain on the index's own scale, where 50 is an average model. Click a domain to see the benchmarks behind it. Pick up to 5 models from the list.
Not drawn, measured for fewer than two of these models: Agents & tools, Maths, Instruction following, Multimodal, Human preference.
ModelIndex
Where sources disagree
Across the 3 benchmarks both have, Claude Fable 5.2 leads 3, Kimi K2.5 leads 0. Every overall index that rates both puts Claude Fable 5.2 ahead.
| Overall index | Claude Fable 5.2 (max) | Kimi K2.5 (thinking) | Leader |
|---|---|---|---|
| BenchLeader Index | 69.4 | 58.8 | Claude Fable 5.2 |
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. Cost, tokens and time per benchmark task are measured: Artificial Analysis’ averages over its whole Intelligence Index run.
| Measure | Claude Fable 5.2 (max) | Kimi K2.5 (thinking) |
|---|---|---|
| Blended price per 1M tokens | $16.00 | $1.20 |
| Input / output per 1M | $8.00 / $40.00 | $0.600 / $3.00 |
| Cached input per 1M | $0.200 | $0.100 |
| Price per task: chat reply | $0.015 | $0.0011 |
| Price per task: summarize a 30-page report | $0.120 | $0.0090 |
| Price per task: code edit | $0.108 | $0.0081 |
| Price per task: agentic coding session | $0.640 | $0.048 |
| Price per task: structured extraction | $0.024 | $0.0018 |
| Agentic coding session, input mostly cached | $0.289 | $0.025 |
| Cost per benchmark task (AA Index) | $26.54 | – |
| Output tokens per task (AA Index) | 274k | – |
| Of which reasoning | 193k | – |
| Output speed | 72 tok/s | 39 tok/s |
| Time to first answer | 3.0 s | 34.1 s |
| Time to a complete answer | – | 105.2 s |
Each model is compared at its best-ranked reasoning-effort setting. Claude Fable 5.2 (max) scores 69.4 and takes 3 s to its first answer token; Claude Fable 5.2 (high) scores 68.9 in 3 s, Claude Fable 5.2 (xhigh) scores 68.7 in 3 s. Kimi K2.5 (thinking) scores 58.8 and takes 34 s to its first answer token; Kimi K2.5 scores 54.3 in 2 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
124 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 Fable 5.2 (max) | Kimi K2.5 (thinking) |
|---|---|---|---|
| BenchLeader Index | BenchLeader | 69.4 | 58.8 |
| Coding score | BenchLeader | 80.2 | 55.6 |
| Knowledge score | BenchLeader | 81.7 | 60.1 |
| Long context score | BenchLeader | 66.4 | 63.6 |
| Reasoning score | BenchLeader | 78.4 | 67.3 |
| Agents & tools score | BenchLeader | – | 45.1 |
| Human preference score | BenchLeader | – | 63.3 |
| Instruction following score | BenchLeader | – | 69.1 |
| Maths score | BenchLeader | – | 56.3 |
| Multimodal score | BenchLeader | – | 59.2 |
| Blended price $/M | BenchLeader | $16.00 | $1.20 |
| Output speed | BenchLeader | 72 tok/s | 39 tok/s |
| Time to first answer | BenchLeader | 3.0 s | 34.1 s |
| Context window | BenchLeader | 1,000,000 tokens | 262,144 tokens |
| Reasoning | |||
| GPQA Diamond | Epoch AI Benchmarking Hub | – | 87.6%† |
| Humanity's Last Exam | Scale AI / CAIS | – | 24.4%† |
| SimpleBench | SimpleBench | – | 46.8%† |
| Arena Hard Prompts | Arena (LMArena) | – | 1471 |
| LiveBench 2.0 Reasoning | LiveBench | 88.8%† | – |
| GPQA Diamond (AA) | Artificial Analysis | – | 87.9% |
| Humanity's Last Exam (AA) | Artificial Analysis | 74.3% | 30.7% |
| HLE-Diamond (AA) | Artificial Analysis | 80.2% | – |
| GPQA Diamond (Vals) | Vals AI | – | 84.1% |
| Kagi LLM Benchmark | Kagi LLM Benchmark | – | 78.5% |
| ARC-AGI-1 | ARC Prize | 97.5% | 65.3%† |
| ARC-AGI-2 | ARC Prize | 92.5% | 11.8%† |
| ARC-AGI-3 | ARC Prize | 94.9% | – |
| CritPt | Artificial Analysis | 35.4% | 3.1% |
| CritPt v2 hard subset (AA) | Artificial Analysis | 76.5% | – |
| MysteryMechanism | Vals AI | 51.8%† | – |
| CL-bench | Epoch AI Benchmarking Hub | – | 19.3%† |
| CL-bench Life | Epoch AI Benchmarking Hub | – | 13.2%† |
| Coding | |||
| SWE-bench Verified (Epoch) | Epoch AI Benchmarking Hub | – | 73.8%† |
| SciCode | SciCode | – | 49.0%† |
| WeirdML | WeirdML | – | 45.6%† |
| Arena Coding | Arena (LMArena) | – | 1502 |
| Arena WebDev | Arena (LMArena) | – | 1436 |
| LiveBench 2.0 Coding | LiveBench | 79.2%† | – |
| SciCode (AA) | Artificial Analysis | 67.7% | – |
Show all 124 benchmarks
| Metric | Publisher | Claude Fable 5.2 (max) | Kimi K2.5 (thinking) |
|---|---|---|---|
| Coding | |||
| LiveCodeBench | Vals AI | – | 83.9% |
| SWE-bench (Vals) | Vals AI | – | 70.0% |
| SWE-bench Verified (bash only) | SWE-bench | – | 70.8%† |
| SWE-bench Verified (any scaffold) | SWE-bench | – | 70.8%† |
| Code Migration | Vals AI | – | 7.0% |
| Code Migration v1.1 | Vals AI | 83.3%† | – |
| ProgramBench | Vals AI | 19.0%† | – |
| SWE-bench Multimodal | Vals AI | 62.3%† | – |
| Vibe Code Bench v1.1 | Vals AI | 93.8%† | 17.5% |
| SWE Atlas: Codebase QnA | Scale AI SEAL | – | 13.1%† |
| SWE Atlas: Refactoring | Scale AI SEAL | – | 20.9%† |
| SWE Atlas: Test Writing | Scale AI SEAL | – | 25.8%† |
| ALE-Bench | Epoch AI Benchmarking Hub | – | 821.6† |
| Terminal-Bench 4.0 (AA) | Artificial Analysis | 59.6% | – |
| Terminal-Bench 2.1 (AA) | Artificial Analysis | – | 45.7% |
| Design Arena Website | Design Arena | – | 1245 |
| FrontierCode 1.1 (AA) | Artificial Analysis | 56.2% | – |
| FrontierCode 2.0 (AA) | Artificial Analysis | 48.8% | – |
| Agents & tools | |||
| Terminal-Bench | Terminal-Bench | – | 43.2%† |
| LiveBench Long Agentic | LiveBench | 82.9%† | – |
| LiveBench Browser | LiveBench | 86.9%† | – |
| Terminal-Bench Hard | Artificial Analysis | – | 34.9% |
| τ²-Bench Telecom (AA) | Artificial Analysis | – | 95.9% |
| Terminal-Bench 2.1 (Vals) | Vals AI | – | 42.0% |
| MCP Atlas | Scale AI SEAL | – | 64.4%† |
| GDPval-AA v2.1 | Artificial Analysis | 1831 | – |
| τ³-Banking (AA) | Artificial Analysis | – | 14.2% |
| APEX-Agents (AA) | Artificial Analysis | – | 11.5% |
| Finance Agent v2 | Vals AI | – | 35.8% |
| Finance Agent v2.1 | Vals AI | 69.5%† | – |
| Harvey's Legal Agent Benchmark v1.1 | Vals AI | 5.0%† | – |
| Harvey's Legal Agent Benchmark v1 | Vals AI | – | 0.0% |
| Legal Research Bench | Vals AI | 46.1%† | 15.9% |
| Tax Agent Bench | Vals AI | 32.1%† | – |
| Terminal-Bench 2.0 (Vals) | Vals AI | – | 40.5% |
| Terminal-Bench 4.0 (Vals) | Vals AI | 65.2%† | – |
| Terminal-Bench Science | Vals AI | 55.7%† | – |
| Vending-Bench 2 | Andon Labs | – | 1198.5† |
| AutomationBench | Artificial Analysis | 76.4% | – |
| GDP.pdf (AA) | Artificial Analysis | 24.2% | – |
| Harvey LAB-AA v1.1 | Harvey | 7.2% | – |
| AA-Briefcase v1.1 | Artificial Analysis | 1785 | – |
| Design Arena Agentic | Design Arena | – | 1111 |
| Terminal-Bench-Science 0.1 (AA) | Artificial Analysis | 73.3% | – |
| Maths | |||
| OTIS Mock AIME | Epoch AI Benchmarking Hub | – | 92.2%† |
| AIME (Vals) | Vals AI | – | 95.6% |
| AIME 2026 | MathArena | – | 95.8% |
| HMMT February 2026 | MathArena | – | 87.1% |
| MathArena Apex | MathArena | – | 8.8% |
| Arena Maths | Arena (LMArena) | – | 1470 |
| Knowledge | |||
| SimpleQA Verified | Epoch AI Benchmarking Hub | – | 34.3%† |
| LiveBench 2.0 Data Analysis | LiveBench | 81.5%† | – |
| LiveBench Research | LiveBench | 94.5%† | – |
| AA-Omniscience v1.1 | Artificial Analysis | 66.8 | – |
| MMLU-Pro | Vals AI | – | 85.9% |
| CorpFin | Vals AI | – | 68.3% |
| TaxEval | Vals AI | – | 74.2% |
| MedQA | Vals AI | – | 94.4% |
| PRBench Finance | Scale AI SEAL | – | 46.5%† |
| PRBench Legal | Scale AI SEAL | – | 43.8%† |
| MultiNRC | Scale AI SEAL | – | 35.2%† |
| BioMysteryBench | Vals AI | 83.7%† | – |
| CaseLaw v2 | Vals AI | – | 58.7% |
| Excel Modeling Benchmark | Vals AI | – | 28.5% |
| Excel Modeling Benchmark v1.1 | Vals AI | 77.4%† | – |
| MedCode | Vals AI | – | 39.3% |
| MedScribe | Vals AI | – | 76.4% |
| AA-Omniscience v1.1: accuracy | Artificial Analysis | 79.1% | – |
| AA-Omniscience v1.1: non-hallucination | Artificial Analysis | 40.7% | – |
| HealthBench Professional (AA) | OpenAI | 67.7% | – |
| Instruction following | |||
| IFBench | Artificial Analysis | – | 70.2% |
| MultiChallenge | Scale AI SEAL | – | 61.4%† |
| TutorBench | Scale AI SEAL | – | 54.6%† |
| Arena Instruction Following | Arena (LMArena) | – | 1439 |
| Human preference | |||
| Arena Text | Arena (LMArena) | – | 1450 |
| Arena Creative Writing | Arena (LMArena) | – | 1422 |
| Arena Multi-turn | Arena (LMArena) | – | 1452 |
| Arena Longer Queries | Arena (LMArena) | – | 1458 |
| Multimodal | |||
| Arena Vision | Arena (LMArena) | – | 1269 |
| LiveBench Design | LiveBench | 85.1%† | – |
| MMMU-Pro | Artificial Analysis | – | 75.4% |
| VISTA | Scale AI SEAL | – | 41.9%† |
| MMMU-Pro (Vals) | Vals AI | – | 84.3% |
| MortgageTax | Vals AI | – | 66.5% |
| AA-3D | Artificial Analysis | 1271 | – |
| Long context | |||
| Fiction.LiveBench 120k | Fiction.live | – | 78.1%† |
| AA-LCR v1.1 | Artificial Analysis | 83.3% | 78.0% |
| Arena Document | Arena (LMArena) | – | 1438 |
| Composite | |||
| Epoch Capabilities Index | Epoch AI Benchmarking Hub | – | 148.7† |
| LiveBench 2.0 | LiveBench | 84.9%† | – |
| AA Intelligence Index v5.0 | Artificial Analysis | 64.6 | – |
| AA Intelligence Index v4.3.2 | Artificial Analysis | – | 23.5 |
| AA Progress Index v1 | Artificial Analysis | – | 270.5 |
| Vals Index | Vals AI | 71.3† | – |
| Vals Finance Index | Vals AI | 63.6%† | – |
| Vals RSI Index | Vals AI | – | 6.8%† |
| Vals Science Index | Vals AI | 63.7%† | – |
| Safety & honesty | |||
| FORTRESS | Scale AI SEAL | – | 41.1%† |
| MASK | Scale AI SEAL | – | 70.5%† |
Data as of 2026-11-09. Best configuration of each model; every score links to its source on the model pages.
What changed
- Kimi K2.5 (thinking): 1 new result (AA Progress Index v1).
- Kimi K2.5 (thinking): 1 new result (Harvey's Legal Agent Benchmark v1).
- Kimi K2.5 (thinking): output price moved from $2.75/M to $3.00/M (9% up).
- Kimi K2.5 (thinking): BenchLeader Index 59.4 → 58.9.
Related comparisons
- Claude Fable 5.2 vs Claude Opus 5.5
- Claude Fable 5.2 vs Gemini 4 Carbon
- Claude Fable 5.2 vs GPT-6 Astra
- Kimi K2.5 vs Claude Opus 5.5
- Kimi K2.5 vs Gemini 4 Carbon
- Kimi K2.5 vs GPT-6 Astra
- Claude Fable 5.2 vs Claude Fable 5.1
- Claude Fable 5.2 vs Claude Fable 5
Rankings: Best LLM overall · Best LLM for agents · Best cheap LLMs
Claude Fable 5.2 vs Kimi K2.5: questions
- Is Claude Fable 5.2 better than Kimi K2.5?
- Claude Fable 5.2 (max) leads on quality: 69.4 vs 58.8. The BenchLeader Index combines every independent quality benchmark; Claude Fable 5.2 (max) is ahead overall as of 9 Nov 2026, but check the category scores for your use.
- Claude Fable 5.2 vs Kimi K2.5: which is cheaper, and will that change?
- Kimi K2.5 is cheaper today: $1.20 against $16.00 per million tokens, blended at three input tokens per output token. No price change has been announced for either.
- Is Claude Fable 5.2 or Kimi K2.5 better for coding?
- Claude Fable 5.2 scores higher in coding (80 vs 56 on the category index, where 50 is average).
- Which is faster, Claude Fable 5.2 or Kimi K2.5?
- Claude Fable 5.2 streams ~1.5× faster (72 vs 39 output tokens per second, Artificial Analysis 7-day median).
- Which is better for long documents?
- Claude Fable 5.2 scores higher on long-context benchmarks (66 vs 64). Claude Fable 5.2 accepts more context: 1,000,000 against 262,144 tokens for Kimi K2.5.
More questions (3)
- Which has lower latency?
- Claude Fable 5.2 answers first: 3.0 s against 34.1 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.
- Do independent benchmarks agree on Claude Fable 5.2 vs Kimi K2.5?
- Across the 3 benchmarks both have, Claude Fable 5.2 leads 3, Kimi K2.5 leads 0. Every overall index that rates both puts Claude Fable 5.2 ahead.
- When was Claude Fable 5.2 released?
- Claude Fable 5.2 was released on 9 Nov 2026.