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. | 1B/4B/12B/27B | 35B |
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. | 128k | 256k |
| 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. | — | $0.375 |
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. | — | $2.25 |
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. | $0.04DeepInfra | $0.05Darkbloom |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $0.05Parasail | $0.70Darkbloom |
| $10.50Parasail · bf16 | $45.00Darkbloom · fp4 | |
| Benchmarks | ||
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 42.4% | 86% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 64.9% | 81.7% |
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. | 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. | — | 49.5% |
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. | — | 73.4% |
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. | — | 67.2% |
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. | 29.7% | — |
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. | 87.8% | — |
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. | — | 51.5% |
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. | — | 21.4% |
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. | 95.9% | — |
CharXiv ReasoningChart reasoning — Can the AI read and reason about complex charts and figures, not just text? Higher is better. | — | 78% |
MMMU-ProMultimodal reasoning — A tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | — | 75.3% |
| Overview | ||
| Company | Qwen | |
| Release date | Mar 12 2025 | Apr 16 2026 |
| Access | Open Weight | Open Weight |
| Model details | View model | View model |
Other comparisons
Gemma 3vsClaude Haiku 5.5Qwen3.6vsClaude Haiku 5.5Gemma 3vsGPT-6.1 SolQwen3.6vsGPT-6.1 SolGemma 3vsMuse Spark OpenQwen3.6vsMuse Spark OpenGemma 3vsGrok 4.7Qwen3.6vsGrok 4.7Gemma 3vsDeepSeek-V4.1-Flash-1017Qwen3.6vsDeepSeek-V4.1-Flash-1017Gemma 3vsMistral Large 4Qwen3.6vsMistral Large 4Frequently asked questions
Qwen3.6 leads Gemma 3 on 2 of the 2 benchmarks they both report (GPQA Diamond, MMMU). Only Qwen3.6 has a verified first-party API price: $0.375 per million input tokens and $2.25 per million output tokens. No pay-as-you-go API rate is tracked for Gemma 3. Gemma 3 shipped 400 days before Qwen3.6, so benchmark comparisons should account for the intervening progress.
Gemma 3 has 1B/4B/12B/27B parameters, while Qwen3.6 has 35B. Context windows are 128k (Gemma 3) vs 256k (Qwen3.6).
On GPQA Diamond, Qwen3.6 leads at 86% vs Gemma 3 at 42.4%. On MMMU, Qwen3.6 leads at 81.7% vs Gemma 3 at 64.9%.