Qwen3 235B-A22B Thinking 2507 (Deep Infra) vs Nemotron 3 Ultra
Qwen3 235B-A22B Thinking 2507 (Deep Infra) is cheaper on output tokens. Choose Qwen3 235B-A22B Thinking 2507 (Deep Infra) or Nemotron 3 Ultra based on the trade-off between cost, context, and the benchmarks that matter for your use case.
| Spec | Qwen3 235B-A22B Thinking 2507 (Deep Infra) | Nemotron 3 Ultra |
|---|---|---|
| Provider | Eden AI | NVIDIA |
| Input / 1M tokens | $0.23 | $0.60 |
| Output / 1M tokens | $2.30 | $2.40 |
| Context window | 262K | 262K |
| Parameters | — | 550B |
| Open weights | Yes | Yes |
| Released | Jul 2025 | Jun 2026 |
| LiveBench score | — | 50.7 |
| LMArena (Chatbot Arena) Elo score | — | 1445 |
| GPQA Diamond score | — | 85.3 |
| Epoch Capabilities Index score | — | 146 |
FAQ
Is Qwen3 235B-A22B Thinking 2507 (Deep Infra) or Nemotron 3 Ultra cheaper?
Qwen3 235B-A22B Thinking 2507 (Deep Infra) is cheaper on output tokens ($2.30 vs $2.40 per 1M).
Which has the larger context window, Qwen3 235B-A22B Thinking 2507 (Deep Infra) or Nemotron 3 Ultra?
Both models offer a similar context window.
Pricing is indicative — confirm with the provider before production use. Highlighted values indicate the better figure for that row.