Head-to-head
Kimi K3 vs Kimi K3.1 Pro: which AI model wins in 2026?
Kimi K3 ($15/1M out) and Kimi K3.1 Pro ($10/1M out) are two of the most-used AI models in 2026. Across 3 community votes, Kimi K3 leads with 57% approval.
Quick verdict
On Reasoning, pick Kimi K3: the arena rates it 4.5/5 against 4/5 for Kimi K3.1 Pro. On budget, Kimi K3.1 Pro wins: it starts at $10/1M out versus $15/1M out for Kimi K3.
Line-by-line comparison
Strengths and weaknesses
Kimi K3
- #3 on the Artificial Analysis Intelligence Index (57) at launch, comparable to Claude Opus 4.8 and GPT-5.5: the closest a Chinese lab has come to the closed US frontier
- 93.4% on SWE-bench Verified in Vals AI's independent harness (GPT-5.6 Sol: 96.2%, Claude Fable 5: 95.0%) and #1 on Arena.ai's Frontend Code Arena at 1679 points, ahead of Fable 5
- 93.5% on GPQA Diamond (between Fable 5's 92.6% and GPT-5.6's 94.1%), 96.1% on AIME 2025, and Elo 1668 on GDPval-AA v2 agentic work, second only to Fable 5
- Strong agentic profile: #1 on AutomationBench-AA SaaS workflows (53%), long-horizon terminal and repo navigation via Kimi Code, and ~21% fewer output tokens than K2 for more intelligence
- $3/$15 per 1M tokens (input drops to $0.30 on cache hit), flat across the full 1M context: still well under US closed-frontier pricing, with an OpenAI-compatible API and OpenRouter availability
- Native multimodal input (text, image, video) and open weights published on Hugging Face under the MIT-style Kimi K3 License, making it the first 3T-class open-weights frontier model
- 3x price jump over Kimi K2.6 ($0.95/$4 to $3/$15) makes it the most expensive Chinese model ever shipped, at Claude Sonnet 5 list price: the '10x cheaper than US models' era is over (Simon Willison documented the hike)
- Slow: 33 output tokens/s, ranked #145 of 190 models on Artificial Analysis, a poor fit for interactive use
- Hallucination rate climbed from 39% (K2.6) to 51% on AA-Omniscience as the model now attempts answers it would previously refuse (Fable 5 sits at a comparable 54.9%)
- Political censorship on sensitive topics in Mandarin (deflects on Xi, CCP, Tiananmen) and Beijing jurisdiction for API data, a compliance blocker for many EU and enterprise deployments; UK AISI/CAISI also found its cyber guardrails failed to block exploit-development attempts during testing
- 'Open weights' is theoretical for most: ~594 GB in native MXFP4 requiring a multi-GPU datacenter to self-host, and the Kimi K3 License demands a separate deal from API resellers above $20M in yearly revenue
Kimi K3.1 Pro
- 65.7% on Terminal-Bench 4.0 per Moonshot, just behind Claude Opus 5.5 (66.4%) and ahead of GPT-6 Astra (58.2%)
- Far fewer hallucinations than K3 on Moonshot's adversarial Hallu Bench Hard: 8.3% of claims wrong with tools vs 23.4% for K3, better than Opus 5.5 (15.0%)
- A third cheaper than Kimi K3: $2/$10 per 1M tokens (cache hit $0.20), batch inference at 60% of list price
- Strong at 3D and spatial work: 92.7 on BenchCAD with tools, level with Claude Fable 5.1 (92.6), and 72.8 without tools, next to Opus 5.5 (73.0)
- Finance and office gains over K3: Financial Research 70.4 vs 58.8, and a state-of-the-art 64.4 on OfficeQA Pro v2 per Moonshot
- Open weights promised by 2026-11-04, after Kimi K3's were published on Hugging Face
- Very slow: 30.6 output tokens per second on Artificial Analysis, #193 of 195; the Highspeed variant doubles the price to $4/$20
- 46 on the Artificial Analysis Intelligence Index v5.0 (#23 of 60), well below Claude Sonnet 5.5 (54) and GPT-6.1 Sol (55) sold at the same $2/$10
- Most headline gains come from Moonshot's internal benchmarks (Kimi Code Bench, Hallu Bench, finance suites) that cannot be checked independently
- Weights and license not yet published at review time, and the API runs under Beijing jurisdiction, a compliance hurdle for many EU and enterprise deployments
- Cyber data was removed from training for the open release, so offensive-security capability sits below frontier closed models (81.0 vs 90.5 for the unfiltered checkpoint on Moonshot's xdev-bench-hard)
Cast your verdict
One recommendation per tool per gladiator. It reshapes the crowd score everyone sees.
The arena’s verdict on Kimi K3
Kimi K3 is the strongest argument yet that the frontier is no longer exclusively American: #3 on Artificial Analysis, top-tier GPQA and SWE-bench Verified scores, and the best frontend-code arena ranking in the business, at roughly half to a third of US closed-frontier prices. Take it for agentic coding, frontend work and SaaS automation where its benchmarks are strongest, or if your roadmap depends on self-hosting a frontier-class model once the weights land. Skip it for interactive products (33 tokens/s is slow), for anything touching politically sensitive content or strict EU data-residency requirements (Mandarin-language censorship, Beijing jurisdiction), and for high-accuracy retrieval where its 51% hallucination rate on AA-Omniscience demands a verification layer. Cost-obsessed teams should note DeepSeek V4 still delivers vastly more tokens per dollar; K3's pitch is peak capability per dollar, not cheapest tokens.
The arena’s verdict on Kimi K3.1 Pro
Kimi K3.1 Pro is a real upgrade over K3 at a lower price: near Opus 5.5 on Moonshot's Terminal-Bench 4.0 figure, standout 3D and CAD skills, and a hallucination rate cut by two thirds on Moonshot's own stress test. Take it for agentic coding, 3D and financial-research work, and especially if you plan to self-host once the weights land. Skip it for anything interactive (30 tokens per second is among the slowest on the market), and compare carefully before buying on price: at the same $2/$10, Claude Sonnet 5.5 and GPT-6.1 Sol score clearly higher on Artificial Analysis and run four to five times faster. K3 users should switch: K3.1 Pro is cheaper and better on every benchmark Moonshot published.
What the crowd says
On Kimi K3
“Tried it for our knowledge-base assistant and the hallucination rate is real: it confidently invented two API endpoints in one afternoon. 33 tokens per second on top of that. Back to waiting for the open weights to fine-tune.”
“Our Zapier-style automation stack runs on it now. Tool orchestration that GPT-5.5 fumbled works reliably, and the cache-hit pricing makes repeated workflows genuinely cheap.”
“The Frontend Arena ranking is deserved. Gave it the same dashboard spec I gave Fable 5 and K3's React came out cleaner, with fewer invented props. At $3/$15 through OpenRouter it was a drop-in swap.”
On Kimi K3.1 Pro
No verdicts yet. Be the first to speak.
Keep comparing
Frequently asked questions
Is Kimi K3 better than Kimi K3.1 Pro?
The crowd currently sides with Kimi K3: 57% recommend it, versus 50% for Kimi K3.1 Pro (3 votes). On Reasoning, Kimi K3 rates higher (4.5/5 vs 4/5). The right pick depends on your use case. The line-by-line comparison on this page breaks down pricing, key specs and arena ratings.
Which is cheaper, Kimi K3 or Kimi K3.1 Pro?
Kimi K3.1 Pro is cheaper: it starts at $10/1M out, while Kimi K3 starts at $15/1M out.
How much do Kimi K3 and Kimi K3.1 Pro cost per 1M tokens?
Kimi K3: $3/1M in per 1M input tokens, $15/1M out per 1M output tokens. Kimi K3.1 Pro: $2/1M in per 1M input tokens, $10/1M out per 1M output tokens.