Compare AI models
| Specifications | ||
Context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider. | 1M | 128k |
| API pricingUSD per 1M tokens · lower wins | ||
Input priceWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it. | $2.00 | $0.15 |
Output priceWhat you pay for the text the model writes back. It is normally the dearer half: producing an answer costs more than reading one. | $10.00 | $0.60 |
Cached input priceA 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. | $0.10 | $0.075 |
Cheapest inputLowest input rate across every provider on OpenRouter, the lab included. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $2.00Amazon Bedrock | $0.15Azure |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $10.00Amazon Bedrock | $0.60Azure |
| $900Anthropic list price | $60.00OpenAI list price | |
| Benchmarks | ||
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. | 60% | 2% |
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. | 81.3% | — |
SWE-Bench MultilingualMultilingual coding — Like SWE-Bench, but the coding problems span many programming languages, not just one. Tests how broadly the AI can code. Higher is better. | 90.3% | — |
SWE-Bench MultimodalMultimodal coding — Real bug reports that arrive with pictures attached — a screenshot, a mockup, a page rendering wrongly — so the AI has to read the image as well as the code to work out what to fix. Higher is better. | 54.3% | — |
CursorBench 4.0Agentic coding — Cursor's own test of coding agents on ambiguous, multi-file tasks taken from real Cursor sessions — editing, refactoring, investigating a codebase, understanding what the user meant, managing jobs and following a design. Cursor runs each model at several reasoning efforts; each release here carries the score of its best listed effort. Scores aren't comparable with earlier CursorBench versions. Higher is better. | 55.5% | — |
FrontierSWE V2Ultra-long-horizon coding — Engineering problems that would occupy a person for days: systems implementation, performance work, scientific computing and AI research, with up to twenty hours per task. Partial credit is awarded, because finishing one outright is still rare. Higher is better. | 61.9% | — |
FrontierCode v1.1 (Main) · main splitAgentic coding — A set of very hard, frontier-difficulty coding tasks an AI agent has to complete end to end. The score is the share of tasks in the main split it solves. Higher is better. | 52.1% | — |
FrontierCode v1.1 (Extended) · extended splitAgentic coding — frontier-difficulty agentic coding tasks (v1.1, extended split) | 64.4% | — |
HumanEvalFunction synthesis — 164 small Python problems: the AI is given a function's description and has to write the working function. This was the coding benchmark of the GPT-3.5 and GPT-4 era, before the field moved to fixing real bugs in real repositories. Higher is better. | — | 87.2% |
Terminal-Bench 4.0Agentic terminal coding — Can the AI work in a command-line terminal — running commands and finishing technical setup tasks the way a developer would? Version 4.0 recalibrated how much time, CPU and memory each task gets, removed eight tasks and fixed nineteen, so fewer runs fail for reasons that have nothing to do with the model. Scores are not comparable with earlier versions. Higher is better. | 70.6% | — |
Terminal-Bench-Science 0.1Agentic scientific computing — The same command-line setup as Terminal-Bench, pointed at scientific work: the AI has to drive research tooling and computational workflows through to a result, rather than administer a machine. Version 0.1 is the first release of the task set, and scores run lower than on the general board. Higher is better. | 59.9% | — |
Humanity's Last Exam · no toolsMultidisciplinary 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. | 56.9% | — |
Humanity's Last Exam · with toolsMultidisciplinary reasoning — Humanity'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. | 64.5% | — |
BioMysteryBench · hardBiology — Real unsolved-style biology puzzles — the AI has to reason its way to an answer the way a research biologist would. The “hard” split contains the toughest cases. Higher is better. | 44.7% | — |
BioMysteryBench · human solvedBiology — Real biology puzzles that human experts have managed to crack — can the AI reach the same answers? Higher is better. | 89.2% | — |
MATHReasoning — accuracy on the original competition mathematics test set | — | 70.2% |
DROPReasoning — three-shot token F1 on questions requiring reasoning over paragraphs | — | 79.7 |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | — | 40.2% |
MMLUGeneral knowledge — A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better. | — | 82% |
MGSMMultilingual math — solving grade-school word problems translated into multiple languages | — | 87% |
OSWorld 2.1 · full setAgentic computer use — An agent operates a desktop across 108 tasks. Each completed checkpoint earns part of the task’s credit, so a model can score even when it only finishes part of a task. | 80.1% | — |
OSWorld 2.1 · offlineAgentic computer use — Can the AI actually operate a computer — clicking, typing, and using real apps — to finish tasks on its own? Version 2.1, in its offline setting. Scores are not comparable with version 2.0. Higher is better. | 83.9% | — |
AutomationBenchBusiness workflows — Tests whether the AI can run real multi-step business workflows — the kind of end-to-end office processes companies want to automate — from start to finish. Higher is better. | 44.7% | — |
HealthBench Professional · rawHealth — physician-graded conversations, scored before adjustment for answer length | 77.1% | — |
HealthBench Professional · length-adjustedHealth — The same physician-graded health conversations as HealthBench Professional, scored with an adjustment for how long the answer is — so a model cannot gain by padding its reply. The adjustment moves scores by several points, so these numbers are not interchangeable with the unadjusted ones. Higher is better. | 69.2% | — |
GDPval-AA v2.1Knowledge work — economically valuable knowledge work (v2.1, Crowd-BT Elo fit) | 1844 | — |
AA-Briefcase v1.1Knowledge work — Artificial Analysis agentic office-work eval (Elo, v1.1 rating fit) | 1811 | — |
Chartography · with toolsChart tasks — A chart-centred test run with tools available to the AI, reported separately from the chart-reading benchmarks above it. Higher is better. | 90.2% | — |
Chartography · no toolsChart tasks — The same chart-centred test with no tools: the AI has to read each chart unaided. Scores run far lower than the with-tools version, so read the two as separate tests. Higher is better. | 61.6% | — |
OfficeQA Pro · extracted textDocument Q&A — answering the 133 Treasury document questions with extracted text and code execution available | 65.6% | — |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | — | 59.4% |
threejsevalCommunity preference (Three.js) — Every model gets the same prompt — "the Eiffel Tower", "a glass fishbowl", "a robot arm picking toys into a box" — and builds a 3D scene in Three.js. Real people then see two scenes side by side, names hidden, and vote for the one they prefer. The votes become a chess-style Elo rating on threejseval.com, averaged across all the prompts. It measures whether the scene looks and moves right to a human eye, not whether the code passes a test. Higher is better. | 1829 | — |
| Overview | ||
| Company | Anthropic | OpenAI |
| Release date | Sep 28 2026 | Jul 18 2024 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Claude Sonnet 5.5vsGemini 4 HydrogenGPT-4o minivsGemini 4 HydrogenClaude Sonnet 5.5vsMuse EmberGPT-4o minivsMuse EmberClaude Sonnet 5.5vsGrok 4.7GPT-4o minivsGrok 4.7Claude Sonnet 5.5vsDeepSeek-V4.1-Flash-1017GPT-4o minivsDeepSeek-V4.1-Flash-1017Claude Sonnet 5.5vsMistral Large 4GPT-4o minivsMistral Large 4Claude Sonnet 5.5vsKimi K3.1 ProGPT-4o minivsKimi K3.1 ProFrequently asked questions
Claude Sonnet 5.5 leads GPT-4o mini on 1 of the 1 benchmark they both report (BullshitBench v2). GPT-4o mini is cheaper on both input and output: $0.15 vs $2.00 per million input tokens, and $0.60 vs $10.00 per million output tokens. GPT-4o mini shipped 802 days before Claude Sonnet 5.5, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Claude Sonnet 5.5) vs 128k (GPT-4o mini).
On BullshitBench v2, Claude Sonnet 5.5 leads at 60% vs GPT-4o mini at 2%.