Gemini 3.1 Pro
Released
Gemini 3.1 Pro is an AI model released by Google on Thursday, Feb 19 2026, 64 days after Gemini 3.0 Flash. Benchmark results (shown below) cover BullshitBench v2, Gray Swan IPI, ProgramBench, SWE-Bench Pro, SWE-Bench Verified, DeepSWE 1.1, and 20 more.
Get Google releases and AI news.
One email, every other Monday.
API pricing
Prompts up to 200K tokens
- InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
- $2.00
- Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
- $0.20
- OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
- $12.00
Prompts over 200K tokens
- InputWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it.
- $4.00
- Cached inputA reduced rate for text you send over and over. If every request starts with the same instructions or the same document, the provider keeps a copy ready and charges less to read it again.
- $0.40
- OutputWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one.
- $18.00
Benchmarks
Coding
ProgramBenchProgram reconstruction — The AI receives a working program and its documentation, then builds a replacement from scratch without the original source code, internet access or decompilation. The score is the percentage of 200 programs that pass every behavioral test. We record each model's best published mini-SWE-agent result, including higher reasoning efforts where available. Partial test-pass rates and almost-solved programs do not count toward this score. Equal scores share a rank here; the official board also uses partial progress to break ties. Higher is better.
0%
via ProgramBench
#5 of 17Best: Claude Opus 5 · 4.5%
SWE-Bench ProAgentic coding — Can 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%
#17 of 23Best: Claude Fable 5.1 · 81.2%
SWE-Bench VerifiedCoding — Real 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%
#10 of 57Best: Claude Opus 5 · 96%
DeepSWE 1.1Agentic coding — Artificial 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%
#25 of 26Best: Muse Spark 1.3 · 75.4%
MLE-BenchML engineering — Can 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%
#3 of 3Best: Gemini 3.6 Flash · 63.9%
Next.js EvalsNext.js coding — Vercel'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.
58%
#19 of 28Best: Claude Fable 5.1 · 97%
Terminal & CLI
Terminal-Bench 2.1Agentic terminal coding — Can 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%
#25 of 30Best: DeepSeek-V4.1-Flash · 90.6%
Terminal-Bench 2.0Agentic terminal coding — Can 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%
#7 of 14Best: Claude Mythos 5 · 88%
Agentic & tool use
MCP AtlasMulti-step tool use — Can 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%
#6 of 11Best: Muse Spark 1.1 · 88.1%
ToolathlonGeneral tool use — Tests how well the AI uses everyday real-world tools and apps to get things done. Higher is better.
48.8%
#5 of 5Best: Gemini 3.5 Flash · 56.5%
BrowseCompWeb browsing — Can the AI browse the web and track down hard-to-find answers? Higher is better.
85.9%
#10 of 28Best: GPT-5.6 Sol · 92.2%
OSWorld-VerifiedAgentic computer use — Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Higher is better.
76.2%
#12 of 28Best: Qwen3.8-Max · 86.1%
Reasoning & science
Humanity's Last ExamMultidisciplinary reasoning — Humanity'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%
no tools
51.4%
with tools
no tools
#5 of 22Best: Claude Fable 5.1 · 60.9%
with tools
#21 of 42Best: Claude Fable 5.1 · 65%
ARC-AGI-2Abstract reasoning — Puzzle-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%
#8 of 21Best: GPT-6 Astra · 95%
FrontierMathAdvanced math — Very hard, research-level math problems. Tiers 1–3 are the (still extremely difficult) lower tiers. Higher is better.
36.9%
Tier 1–3
16.7%
Tier 4
Tier 1–3
#6 of 6Best: GPT-5.5-Pro · 52.4%
Tier 4
#6 of 6Best: GPT-5.5-Pro · 39.6%
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better.
94.3%
#3 of 59Best: GPT-6 Astra · 96%
Knowledge work
GDPval-AAKnowledge work — Measures 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
#8 of 9Best: Claude Fable 5 · 1932
GDPval-AA v2Knowledge work — economically valuable knowledge work (v2, re-based Elo)
965
#19 of 20Best: Claude Opus 5 · 1861
GDPval (win/tie rate)Knowledge work — How often the AI's work matches or beats a human expert's on real knowledge-work tasks. Higher is better.
67.3%
#6 of 6Best: GPT-5.5 · 84.9%
Finance
Finance Agent v2Agentic financial analysis — Tests the AI on real financial-analysis work, like digging through reports and making sound decisions. Higher is better.
43%
#8 of 9Best: Gemini 3.8 Flash · 61.4%
Multimodal
CharXiv ReasoningChart reasoning — Can the AI read and reason about complex charts and figures, not just text? Higher is better.
83.3%
#10 of 16Best: Muse Spark · 88.9%
MMMU-ProMultimodal reasoning — A tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better.
80.5%
#7 of 12Best: Gemini 3.5 Flash · 83.6%
Blueprint-Bench 2Spatial reasoning — Can the AI reason about space and layout — for example, understanding a floor plan or blueprint? Higher is better.
26.5%
#3 of 6Best: GPT-5.5 · 36.2%
Long context
MRCR v2 (8-needle)Long context — Tests 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%
128k average
26.3%
1M pointwise
128k average
#3 of 7Best: Gemini 3.7 Flash · 97%
1M pointwise
#3 of 4Best: Gemini 3.7 Flash · 62.5%
Robustness
BullshitBench v2Nonsense detection — Given 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%
#41 of 77Best: Claude Opus 4.8 · 95%
Gray Swan IPIPrompt injection robustness — Attackers 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%
k = 1
45.7%
k = 10
49.2%
k = 15
k = 1
#13 of 13Best: Claude Opus 5 · 0.2%
k = 10
#11 of 13Best: Claude Opus 5 · 1.6%
k = 15
#11 of 13Best: Claude Opus 5 · 2%
Source: ProgramBench, retrieved 11 September 2026. Every other score here is the figure the lab published at launch.
About
Gemini 3.1 Pro, released February 19, 2026, sharpened the 3-series flagship into a reasoning specialist: 94.3% on GPQA Diamond and 77.1% on ARC-AGI-2 were both the top published scores at release, alongside 85.9% on BrowseComp for agentic web research and 84.9% on MRCR v2 128K long-context recall.
It landed in the same fortnight as Anthropic's Claude Sonnet 4.6 and xAI's Grok 4.20 Beta, in one of 2026's densest release windows. The 3.1 generation filled out below it with Gemini 3.1 Flash-Lite in early March, before Google's attention shifted to the 3.5 Flash line in May 2026.
Compare Gemini 3.1 Pro with
Suggested comparisons
Gemini 3.1 ProvsGemini 3.8 FlashGemini 3.1 ProvsGPT-6 AstraGemini 3.1 ProvsClaude Fable 5.1Gemini 3.1 ProvsMuse Spark 1.3Gemini 3.1 ProvsGrok 4.6Gemini 3.1 ProvsDeepSeek-V4.1-FlashGemini 3.1 ProvsMistral Medium 3.5Gemini 3.1 ProvsKimi K3Gemini 3.1 ProvsGLM-5.3-FlashGemini 3.1 ProvsQwen3.8-Max-0902Gemini 3.1 ProvsNemotron 3.5 Lightning
Frequently asked questions
Gemini 3.1 Pro was released by Google on Thursday, Feb 19 2026.