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 | 1M |
| 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. | — | $0.40 |
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. | — | $1.60 |
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 |
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. | — | $0.40Azure |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | — | $1.60Azure |
| — | $160OpenAI list price | |
| Benchmarks | ||
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. | 93.9% | 23.6% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 94.5% | 65% |
CharXiv ReasoningChart reasoning — Can the AI read and reason about complex charts and figures, not just text? Higher is better. | 86.1% | 56.8% |
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. | 77.8% | — |
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. | 87.3% | — |
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. | 59% | — |
Aider Polyglot (diff)Coding — editing code in six languages using the diff edit format | — | 31.6% |
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. | — | 93.8% |
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. | 82% | — |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 84.1% |
CyberGymCybersecurity — Tests the AI on cybersecurity challenges — finding and exploiting software weaknesses inside a safe sandbox. Higher is better. | 83.1% | — |
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. | 56.8% | — |
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. | 64.7% | — |
MATH-500Reasoning — accuracy on the 500-problem MATH subset | — | 81.4% |
DROPReasoning — three-shot token F1 on questions requiring reasoning over paragraphs | — | 81 |
AIME 2024Reasoning — single-answer accuracy on the 2024 invitational mathematics exam without tools | — | 49.6% |
SimpleQAKnowledge — accuracy on short, fact-seeking questions in the original SimpleQA set | — | 16.8% |
MMMLUKnowledge — average accuracy on multilingual translations of MMLU | — | 78.5% |
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. | — | 87.5% |
MGSMMultilingual math — solving grade-school word problems translated into multiple languages | — | 88.2% |
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. | 79.6% | — |
CharXiv Reasoning · with toolsChart reasoning — The same chart-and-figure reasoning test, run with the AI allowed to use tools — writing code to inspect the image, for instance — rather than reading the chart unaided. Scores run higher than the unaided version, so read the two as separate tests. Higher is better. | 93.2% | — |
MathVistaMultimodal reasoning — mathematical reasoning about visual information | — | 73.1% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | — | 72.7% |
GraphWalks BFS (<128K)Long context — breadth-first traversal accuracy for inputs shorter than 128K tokens | — | 61.7% |
GraphWalks parents (<128K)Long context — parent-node retrieval accuracy for inputs shorter than 128K tokens | — | 60.5% |
OpenAI MRCR (2 needles) · 128KLong context — retrieval of two matching turns in the original MRCR evaluation | — | 47.2% |
OpenAI MRCR (2 needles) · 1MLong context — retrieval of two matching turns in the original MRCR evaluation | — | 33.3% |
GraphWalks BFS (256K–1M)Long context — breadth-first search F1 on the 200 GraphWalks problems with 256K to 1M tokens of context | 80% | — |
| Overview | ||
| Company | Anthropic | OpenAI |
| Release date | Apr 7 2026 | Apr 14 2025 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Claude Mythos PreviewvsGemini 4 CarbonGPT-4.1 minivsGemini 4 CarbonClaude Mythos PreviewvsMuse EmberGPT-4.1 minivsMuse EmberClaude Mythos PreviewvsGrok 4.7GPT-4.1 minivsGrok 4.7Claude Mythos PreviewvsDeepSeek-V4.1-Flash-1017GPT-4.1 minivsDeepSeek-V4.1-Flash-1017Claude Mythos PreviewvsMistral Large 4GPT-4.1 minivsMistral Large 4Claude Mythos PreviewvsKimi K3.1 ProGPT-4.1 minivsKimi K3.1 ProFrequently asked questions
Claude Mythos Preview leads GPT-4.1 mini on 3 of the 3 benchmarks they both report (SWE-Bench Verified, GPQA Diamond, CharXiv Reasoning). Only GPT-4.1 mini has a verified first-party price for an available API: $0.40 per million input tokens and $1.60 per million output tokens. No currently available pay-as-you-go API rate is tracked for Claude Mythos Preview. GPT-4.1 mini shipped 358 days before Claude Mythos Preview, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Claude Mythos Preview) vs 1M (GPT-4.1 mini).
On SWE-Bench Verified, Claude Mythos Preview leads at 93.9% vs GPT-4.1 mini at 23.6%. On GPQA Diamond, Claude Mythos Preview leads at 94.5% vs GPT-4.1 mini at 65%. On CharXiv Reasoning, Claude Mythos Preview leads at 86.1% vs GPT-4.1 mini at 56.8%.