Kimi K3 vs GPT-5.5
Moonshot AI Kimi K3 | OpenAI GPT-5.5 | |
|---|---|---|
| Overview | ||
| Company | Moonshot AI | OpenAI |
| Release date | Jul 16 2026 | Apr 23 2026 |
| Access | Open Weight | Proprietary |
| Specifications | ||
Parameters | 2.8T | — |
Context window | 1M | 1.05M |
| Benchmarks | ||
Nonsense detection BullshitBench v2Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | — | 47% |
Agentic coding SWE-Bench ProCan the AI fix real bugs in real software? It's handed actual problems from open-source projects and has to write code that genuinely solves them. Higher is better. | — | 58.6% |
Multilingual coding SWE-Bench MultilingualLike SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | — | 77.8% |
Agentic coding CursorBench v3.1Cursor's own test of harder, real-world coding tasks inside a code editor. Higher is better. | — | 64.3% |
Agentic coding DeepSWE 1.0Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. Higher is better. | 67.5%Best | 64.3% |
Agentic terminal coding Terminal-Bench 2.1Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Higher is better. | 88.3%Best | 78.2% |
Agentic terminal coding Terminal-Bench 2.0Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? (Version 2.0 of the test.) Higher is better. | — | 82.7% |
Software engineering Expert-SWE (Internal)OpenAI's private set of expert-level software-engineering problems. Higher is better. | — | 73.1% |
Multi-step tool use MCP AtlasCan the AI chain together many tools and steps to complete one bigger task, rather than doing just a single thing? Higher is better. | 84.2%Best | 75.3% |
Professional tool use JobBenchTests the AI on professional workplace tasks that require using real work tools — the kind of multi-step jobs an office worker handles. Higher is better. | 52.9% | — |
Personal tool use Toolathlon-VerifiedTests how well the AI uses everyday personal tools and apps to get things done — a human-checked version of Toolathlon. Higher is better. | 73.2% | — |
General tool use ToolathlonTests how well the AI uses everyday real-world tools and apps to get things done. Higher is better. | — | 55.6% |
Web browsing BrowseCompCan the AI browse the web and track down hard-to-find answers? Higher is better. | 91.2%Best | 84.4% |
Cybersecurity CyberGymTests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better. | — | 81.8% |
Multidisciplinary reasoning Humanity's Last Exam · no toolsHumanity's Last Exam — extremely hard expert questions across many subjects, written so you can't just look up the answer. “No tools” means the AI answers on its own. Higher is better. | 43.5%Best | 41.4% |
Multidisciplinary reasoning Humanity's Last Exam · with toolsHumanity's Last Exam — extremely hard expert questions across many subjects. “With tools” means the AI is allowed to search the web or run code while answering. Higher is better. | 56%Best | 52.2% |
Abstract reasoning ARC-AGI-2Puzzle-style tests of abstract reasoning and pattern-finding — the kind of thing people find easy but AIs often struggle with. Higher is better. | — | 84.6% |
Advanced math FrontierMath · Tier 1–3Very hard, research-level math problems. Tiers 1–3 are the (still extremely difficult) lower tiers. Higher is better. | — | 51.7% |
Advanced math FrontierMath · Tier 4Very hard, research-level math problems. Tier 4 is the hardest — close to what professional research mathematicians tackle. Higher is better. | — | 35.4% |
Science GPQA DiamondGraduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 93.5% | 93.6%Best |
Agentic computer use OSWorld-VerifiedCan the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better. | — | 78.7% |
Agentic financial analysis Finance Agent v2Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. | — | 51.8% |
Agentic legal work Harvey's Legal Agent BenchmarkHarvey's test of whether an AI agent can complete real legal work — drafting and reviewing documents, working with spreadsheets and presentations, and navigating files the way a lawyer's assistant would. Higher is better. | — | 3.75% |
Tax questions TaxEval v2A set of real tax questions created by Vals AI — can the AI give accurate answers about tax rules and filings? Higher is better. | — | 74.98% |
Medical admin work MedScribeCan the AI support doctors with their administrative work, like notes and paperwork? Created by Vals AI. Higher is better. | — | 86.87% |
Knowledge work GDPval-AAMeasures how well the AI does economically valuable knowledge work, judged against human experts. Shown as a rating (like a chess Elo) — higher is better. | — | 1769 |
Knowledge work GDPval-AA v2economically valuable knowledge work (v2, re-based Elo) | 1668Best | 1494 |
Knowledge work GDPval (win/tie rate)How often the AI's work matches or beats a human expert's on real knowledge-work tasks. Higher is better. | — | 84.9% |
Chart reasoning CharXiv ReasoningCan the AI read and reason about complex charts and figures, not just text? Higher is better. | 84.8%Best | 84.1% |
Multimodal reasoning MMMU-ProA tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 81.6%Best | 81.2% |
Spatial reasoning Blueprint-Bench 2Can the AI reason about space and layout — for example, understanding a floor plan or blueprint? Higher is better. | — | 36.2% |
Long context MRCR v2 (8-needle) · 128k averageTests whether the AI can find specific details buried inside a very long document (around 128k tokens — roughly a long book). Higher is better. | — | 94.8% |
Community preference Arena Elo (Text)Real people chat with two anonymous AIs side by side and vote for the answer they prefer. Votes become a chess-style Elo rating on arena.ai — it measures which AI people actually like, not test scores. Higher is better. | — | 1481 |
Community preference (code) Arena Elo (Code)Like the text arena, but people vote on which AI writes better code. The votes become a chess-style Elo rating on arena.ai. Higher is better. | — | 1502 |
| Timeline | ||
| Release gap | GPT-5.5 shipped 84 days before Kimi K3 | |
Which is better: Kimi K3 or GPT-5.5?
Kimi K3 leads GPT-5.5 on 9 of the 10 benchmarks they both report. GPT-5.5 shipped 84 days before Kimi K3, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Kimi K3) vs 1.05M (GPT-5.5). Kimi K3 is open weight, while GPT-5.5 is proprietary.
On DeepSWE 1.0, Kimi K3 leads at 67.5% vs GPT-5.5 at 64.3%. On Terminal-Bench 2.1, Kimi K3 leads at 88.3% vs GPT-5.5 at 78.2%. On MCP Atlas, Kimi K3 leads at 84.2% vs GPT-5.5 at 75.3%. On BrowseComp, Kimi K3 leads at 91.2% vs GPT-5.5 at 84.4%. On Humanity's Last Exam · no tools, Kimi K3 leads at 43.5% vs GPT-5.5 at 41.4%. On Humanity's Last Exam · with tools, Kimi K3 leads at 56% vs GPT-5.5 at 52.2%. On GPQA Diamond, GPT-5.5 leads at 93.6% vs Kimi K3 at 93.5%. On GDPval-AA v2, Kimi K3 leads at 1668 vs GPT-5.5 at 1494. On CharXiv Reasoning, Kimi K3 leads at 84.8% vs GPT-5.5 at 84.1%. On MMMU-Pro, Kimi K3 leads at 81.6% vs GPT-5.5 at 81.2%.
Frequently asked questions
Kimi K3 was released by Moonshot AI on Jul 16 2026.
GPT-5.5 was released by OpenAI on Apr 23 2026.
Kimi K3 leads on DeepSWE 1.0 — Kimi K3 67.5% vs GPT-5.5 64.3%.
Kimi K3 leads on Humanity's Last Exam · no tools — Kimi K3 43.5% vs GPT-5.5 41.4%.
Kimi K3 has a 1M context window; GPT-5.5 has 1.05M.
Kimi K3 is an open weight model released by Moonshot AI. GPT-5.5 is a proprietary model released by OpenAI.