Compare AI models
| Specifications | ||
ParametersThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. | — | 2.4T |
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. | 1.05M | 1M |
| API pricingUSD per 1M tokens · lower wins · base tier | ||
Input priceWhat you pay for everything you send the model — your question, plus any documents or earlier conversation you include with it. | $1.00 | — |
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. | $5.00 | — |
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. | $0.05 | — |
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. | — | $2.00Alibaba |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | — | $6.00Alibaba |
| $450OpenAI list price | $700Alibaba | |
| Benchmarks | ||
MMMU-ProMultimodal reasoning — A tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 85.5% | 82.3% |
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. | — | 95% |
OpenAI Factuality (Internal)Factual error rate — OpenAI's internal test uses prompts designed to provoke factual mistakes. The score is the percentage of answers containing at least one factual error, at the model's best published reasoning effort. It is not the percentage of individual claims that are wrong or an error rate for everyday use. Lower is better. | 4.3% | — |
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. | — | 67.7% |
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. | — | 56.6% |
PaperBenchResearch reproduction — reproducing the results of an ML research paper end to end | — | 93% |
NL2Repo-BenchRepo-level code generation — Tests whether the AI can turn a natural-language requirement into working code across an entire repository, not just produce a single function or patch. Higher is better. | — | 55.9% |
QwenSWEBench V2Software engineering — Qwen's in-house coding benchmark, second version, built around complex real-world software-engineering tasks. Scores are not comparable with the first version. Higher is better. | — | 55.1% |
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. | 50.3% | — |
Terminal-Bench 3.0Agentic terminal coding — command-line task completion (v3.0, much harder task set) | — | 11.3% |
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. | — | 86.6% |
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. | 58.3% | — |
JobBenchProfessional tool use — Tests the AI on professional workplace tasks that require using real work tools — the kind of multi-step jobs an office worker handles. Higher is better. | — | 53.4% |
CoWorkBenchLong-horizon office work — Tests long-running office tasks across fields including computer science, finance, law, medicine, and other productivity work. Higher is better. | — | 74.8% |
Toolathlon-VerifiedPersonal tool use — Tests how well the AI uses everyday personal tools and apps to get things done — a human-checked version of Toolathlon. Higher is better. | — | 72.5% |
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. | — | 43.6% |
OSWorld 2.1 (offline)Agentic 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. | 74.2% | — |
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. | — | 86.1% |
Agent's Last Exam · pass@1Agentic computer use — A hard set of desktop and operating-system tasks an AI agent has to finish by looking at the screen and working the machine itself. The score is the share it passes outright — partial credit does not count. Higher is better. | 60% | — |
CharXiv ReasoningChart reasoning — Can the AI read and reason about complex charts and figures, not just text? Higher is better. | — | 88.4% |
BabyVisionVisual reasoning — Tests core visual reasoning — seeing and understanding images the way even young children can, which AIs often find surprisingly hard. Higher is better. | — | 82% |
| Overview | ||
| Company | OpenAI | Qwen |
| Release date | Oct 26 2026 | Aug 3 2026 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
GPT-6.1 TerravsClaude Haiku 5.5Qwen3.8-MaxvsClaude Haiku 5.5GPT-6.1 TerravsGemini 4 ArgonQwen3.8-MaxvsGemini 4 ArgonGPT-6.1 TerravsMuse Spark OpenQwen3.8-MaxvsMuse Spark OpenGPT-6.1 TerravsGrok 4.7Qwen3.8-MaxvsGrok 4.7GPT-6.1 TerravsDeepSeek-V4.1-Flash-1017Qwen3.8-MaxvsDeepSeek-V4.1-Flash-1017GPT-6.1 TerravsMistral Large 4Qwen3.8-MaxvsMistral Large 4Frequently asked questions
GPT-6.1 Terra leads Qwen3.8-Max on 1 of the 1 benchmark they both report (MMMU-Pro). Only GPT-6.1 Terra has a verified first-party API price: $1.00 per million input tokens and $5.00 per million output tokens. No pay-as-you-go API rate is tracked for Qwen3.8-Max. Qwen3.8-Max shipped 84 days before GPT-6.1 Terra, so benchmark comparisons should account for the intervening progress.
Context windows are 1.05M (GPT-6.1 Terra) vs 1M (Qwen3.8-Max).
On MMMU-Pro, GPT-6.1 Terra leads at 85.5% vs Qwen3.8-Max at 82.3%.