ModelScalePreferred languages
More languages supported by Google Translate
Open the selector to load the current language catalog.
Translations are provided by Google Translate.
Compare
Two to four models, side by side. The link preserves the canonical model ids, and every value that a source did not return stays explicitly unavailable.
Reading the pair evidence for the models in this link.
Check two to four models below, then press Compare — the URL and evidence below only update once you do.
Exact source values, missing states, and pair-derived deltas stay separate.
Evidence key: ObservedUnavailable
| Capability | Hy3 | Qwen3.7 Flash |
|---|---|---|
| Agentic | Unavailable | Unavailable |
| Coding | Unavailable | Unavailable |
| Knowledge | 57.7 | Unavailable |
| Reasoning | 75.3 | Unavailable |
| Multimodal & Grounded | Unavailable | Unavailable |
| Instruction Following | Unavailable | Unavailable |
| Math | Unavailable | Unavailable |
| Overall | 60.79 | 49.46 |
Time to first token and throughput, exactly as the runtime source published them.
Time to first token: latency to the first answer chunk (Artificial Analysis, via BenchLM). Reasoning models include thinking time, so values can run to tens or hundreds of seconds.
Source rates plus a labelled 75% input / 25% output blended cost.
| Metric | Hy3 Self-hosted | Qwen3.7 Flash |
|---|---|---|
| Input / 1M | $0.0825 | $0.03 |
| Output / 1M | $0.33 | $0.13 |
| Cache read / 1M | $0.0206 | Unavailable |
| Cache write / 1M | Unavailable | Unavailable |
| Blended / 1M (75% input / 25% output) | $0.14 | $0.055 |
| Cost per successful task (LiveBench) | Unavailable | Unavailable |
| Time to first token | 28.33 s | Unavailable |
| Throughput | 78 tok/s | Unavailable |
| Context window | 256K tokens | 1M tokens |
| Release date | 2026-07-06 | 2026-07-27 |
| Access | Open Weight | Proprietary |
1 of these rows is marked self-hosted by the source: open-weight models with no first-party hosted token rate. 1 of them shows a hosted rate matched from another provider; the rest report an unavailable price rather than zero.
Each row comes from the pair comparison source. A null basis or average stays unavailable rather than being inferred from the capability table.
The comparison source only publishes deltas for pairs it has evaluated. The capability and price tables above remain source-exact.