Gemini 3.1 ProvsComposer 2.5
Gemini 3.1 Pro | Composer 2.5 | |
|---|---|---|
| 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. | 37% | — |
Prompt injection robustness Gray Swan IPI · k = 1Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. Gray Swan's indirect prompt injection benchmark measures how often such an attack succeeds when the attacker gets a single try. Lower is better. | 14.2% | — |
Prompt injection robustness Gray Swan IPI · k = 10Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 10 tries and counts an attack as successful if any of them works. Lower is better. | 45.7% | — |
Prompt injection robustness Gray Swan IPI · k = 15Attackers hide malicious instructions inside content the AI reads — a web page, an email, a document — and try to hijack what it does. This variant gives the attacker 15 tries and counts an attack as successful if any of them works. Lower is better. | 49.2% | — |
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. | 54.2%Best | 54% |
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. | 80.6% | — |
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. | — | 79.8% |
Agentic coding CursorBench v3.2Cursor's own test of harder, real-world coding tasks inside a code editor, on the refreshed v3.2 task set. Scores aren't comparable with v3.1. Higher is better. | — | 56.1% |
Agentic coding CursorBench v3.1Cursor's own test of harder, real-world coding tasks inside a code editor. Higher is better. | — | 63.2% |
Agentic coding DeepSWE 1.1Artificial Analysis' independent test of deep, agentic software-engineering work — the AI has to plan and carry out substantial coding tasks end to end. (Version 1.1 of the test.) Higher is better. | 12% | — |
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. | — | 18% |
ML engineering MLE-BenchCan the AI do the work of a machine-learning engineer? It competes in real Kaggle competitions — building, training, and tuning models end to end — and the score reflects how well it places. Higher is better. | 42.6% | — |
Next.js coding Next.js EvalsVercel's open eval of how well AI coding agents build and migrate real Next.js apps — measured as the share of tasks the agent completes successfully. Higher is better. | 75% | 92%Best |
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. | 70.3% | 73%Best |
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. | 68.5% | 69.3%Best |
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. | 78.2% | — |
General tool use ToolathlonTests how well the AI uses everyday real-world tools and apps to get things done. Higher is better. | 48.8% | — |
Web browsing BrowseCompCan the AI browse the web and track down hard-to-find answers? Higher is better. | 85.9% | — |
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. | 44.4% | — |
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. | 51.4% | — |
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. | 77.1% | — |
Advanced math FrontierMath · Tier 1–3Very hard, research-level math problems. Tiers 1–3 are the (still extremely difficult) lower tiers. Higher is better. | 36.9% | — |
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. | 16.7% | — |
Science GPQA DiamondGraduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 94.3% | — |
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. | 76.2% | — |
Agentic financial analysis Finance Agent v2Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better. | 43% | — |
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. | 1314 | — |
Knowledge work GDPval-AA v2economically valuable knowledge work (v2, re-based Elo) | 965 | — |
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. | 67.3% | — |
Chart reasoning CharXiv ReasoningCan the AI read and reason about complex charts and figures, not just text? Higher is better. | 83.3% | — |
Multimodal reasoning MMMU-ProA tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 80.5% | — |
Spatial reasoning Blueprint-Bench 2Can the AI reason about space and layout — for example, understanding a floor plan or blueprint? Higher is better. | 26.5% | — |
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. | 84.9% | — |
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. | 26.3% | — |
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. | 1485 | — |
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. | 1445 | — |
| Overview | ||
| Company | SpaceXAI | |
| Release date | Feb 19 2026 | May 18 2026 |
| Access | Proprietary | Proprietary |
Which is better: Gemini 3.1 Pro or Composer 2.5?
Composer 2.5 leads Gemini 3.1 Pro on 3 of the 4 benchmarks they both report (SWE-Bench Pro, Next.js Evals, Terminal-Bench 2.1, Terminal-Bench 2.0). Gemini 3.1 Pro shipped 88 days before Composer 2.5, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On SWE-Bench Pro, Gemini 3.1 Pro leads at 54.2% vs Composer 2.5 at 54%. On Next.js Evals, Composer 2.5 leads at 92% vs Gemini 3.1 Pro at 75%. On Terminal-Bench 2.1, Composer 2.5 leads at 73% vs Gemini 3.1 Pro at 70.3%. On Terminal-Bench 2.0, Composer 2.5 leads at 69.3% vs Gemini 3.1 Pro at 68.5%.
Frequently asked questions
Gemini 3.1 Pro was released by Google on Feb 19 2026.
Composer 2.5 was released by SpaceXAI on May 18 2026.
Gemini 3.1 Pro leads on SWE-Bench Pro — Gemini 3.1 Pro 54.2% vs Composer 2.5 54%.
Other comparisons
Gemini 3.1 ProvsClaude Opus 5Composer 2.5vsClaude Opus 5Gemini 3.1 ProvsGPT-5.6-CyberComposer 2.5vsGPT-5.6-CyberGemini 3.1 ProvsMuse GlimmerComposer 2.5vsMuse GlimmerGemini 3.1 ProvsDeepSeek-V4-Pro-0813Composer 2.5vsDeepSeek-V4-Pro-0813Gemini 3.1 ProvsMistral Medium 3.5Composer 2.5vsMistral Medium 3.5Gemini 3.1 ProvsKimi K3Composer 2.5vsKimi K3