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. | 7B/14B | — |
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. | — | $2.50International (Singapore) |
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. | — | $7.50 |
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. | — | $1.475Alibaba |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | — | $4.425Alibaba |
| — | $516Alibaba | |
| Benchmarks | ||
LiveCodeBenchCompetitive coding — Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better. | 38.6% | 91.6% |
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. | — | 72% |
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. | — | 60.6% |
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. | — | 80.4% |
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. | — | 78.3% |
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. | 88.4% | — |
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. | — | 69.7% |
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. | — | 76.4% |
IFEvalAgentic tasks — following verifiable instructions in generated responses | 84.3% | — |
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. | — | 41.4% |
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. | — | 53.5% |
MATHReasoning — accuracy on the original competition mathematics test set | 79.5% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | — | 92.4% |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 63.3% | — |
GSM8KGrade-school math — Grade-school maths word problems that take a few steps of arithmetic to work through. It separated the models of 2022 and 2023 sharply, then saturated. One caveat on the historical numbers: OpenAI included part of the GSM8K training set in GPT-4's pre-training mix, so GPT-4's score is not a clean few-shot result. Higher is better. | 94.8% | — |
| Overview | ||
| Company | Qwen | Qwen |
| Release date | Jan 27 2025 | May 20 2026 |
| Access | Open Weight | Closed |
| Model details | View model | View model |
Other comparisons
Qwen2.5-1MvsClaude Fable 5.2Qwen3.7-MaxvsClaude Fable 5.2Qwen2.5-1MvsGPT-6.2 AstraQwen3.7-MaxvsGPT-6.2 AstraQwen2.5-1MvsGemini 4 NeonQwen3.7-MaxvsGemini 4 NeonQwen2.5-1MvsMuse EmberQwen3.7-MaxvsMuse EmberQwen2.5-1MvsGrok 4.7Qwen3.7-MaxvsGrok 4.7Qwen2.5-1MvsDeepSeek-V4.1-Flash-1017Qwen3.7-MaxvsDeepSeek-V4.1-Flash-1017Frequently asked questions
Qwen3.7-Max leads Qwen2.5-1M on 1 of the 1 benchmark they both report (LiveCodeBench). Only Qwen3.7-Max has a verified first-party price for an available API: $2.50 per million input tokens and $7.50 per million output tokens. No currently available pay-as-you-go API rate is tracked for Qwen2.5-1M. The listed Qwen3.7-Max rates use the “International (Singapore)” pricing scope. Qwen2.5-1M shipped 478 days before Qwen3.7-Max, so benchmark comparisons should account for the intervening progress.
Context windows are 1M (Qwen2.5-1M) vs 1M (Qwen3.7-Max). Qwen2.5-1M is open weight, while Qwen3.7-Max is closed.
On LiveCodeBench, Qwen3.7-Max leads at 91.6% vs Qwen2.5-1M at 38.6%.