1 vs 0 benchmarks won
Google Gemini 3.0 Flash | Moonshot AI Kimi K2 Thinking | |
|---|---|---|
| Overview | ||
| Company | Moonshot AI | |
| Release date | Dec 17 2025 | Nov 6 2025 |
| Model type | — | — |
| Open source | No | Yes |
| Specifications | ||
Parameters | — | 1T |
Context window | — | 256k |
| Benchmarks | ||
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. | 49.6% | — |
Coding SWE-Bench VerifiedReal coding tasks pulled from open-source projects — the AI has to find and fix actual bugs. A human-checked version of the original SWE-Bench. Higher is better. | 78%Best | 71.3% |
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. | 58% | — |
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. | 62% | — |
General tool use ToolathlonTests how well the AI uses everyday real-world tools and apps to get things done. Higher is better. | 49.4% | — |
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. | 33.7% | — |
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. | 33.6% | — |
Science GPQA DiamondGraduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 90.4% | — |
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. | 65.1% | — |
Agentic financial analysis Finance Agent v2Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. | 42.6% | — |
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. | 1204 | — |
Chart reasoning CharXiv ReasoningCan the AI read and reason about complex charts and figures, not just text? Higher is better. | 80.3% | — |
Multimodal reasoning MMMU-ProA tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 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. | 0% | — |
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. | 67.2% | — |
Long context MRCR v2 (8-needle) · 1M pointwiseTests whether the AI can find specific details buried inside an enormous document (around 1 million tokens — many books). Higher is better. | 22.1% | — |
| Timeline | ||
| Release gap | Kimi K2 Thinking shipped 41 days before Gemini 3.0 Flash | |
Gemini 3.0 Flash leads Kimi K2 Thinking on 1 of the 1 benchmark they both report (SWE-Bench Verified). Kimi K2 Thinking shipped 41 days before Gemini 3.0 Flash, so benchmark comparisons should account for the intervening progress.
Kimi K2 Thinking is an open-source / open-weight model; Gemini 3.0 Flash is proprietary.
On SWE-Bench Verified, Gemini 3.0 Flash leads at 78% vs Kimi K2 Thinking at 71.3%.