Nemotron 3 Ultra vs Qwen3 235B A22B Thinking 2507 API pricing
Updated
Nemotron 3 Ultra: USD 0.50 input, USD 2.15 output per 1M tokens (CoreWeave Forge). Qwen3 235B A22B Thinking 2507: USD 0.45 input, USD 3.50 output per 1M tokens (Featherless and 1 other). At 3 input tokens per output token, Qwen3 235B A22B Thinking 2507 costs 33% more than Nemotron 3 Ultra (USD 1.212 vs USD 0.9125 per 1M).
How do Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 differ on price?
- Cheapest input price: Qwen3 235B A22B Thinking 2507 USD 0.35 against USD 0.50 per 1M tokens, so Nemotron 3 Ultra costs 1.43× as much (43% more).
- Cheapest output price: Nemotron 3 Ultra USD 2.15 against USD 3.50 per 1M tokens, so Qwen3 235B A22B Thinking 2507 costs 1.63× as much (63% more).
- Among the 201 indexed models, Nemotron 3 Ultra ranks 128th cheapest and Qwen3 235B A22B Thinking 2507 138th on the 3:1 blend.
- The cheaper model changes at 14 input tokens per output token.
- Only Nemotron 3 Ultra lists a cached-input discount: Nemotron 3 Ultra at 10 standard prices of 10, from USD 0.10 (CoreWeave Forge, 80% below that provider's input price). Qwen3 235B A22B Thinking 2507's cached tokens bill at the input price.
- Nemotron 3 Ultra states up to 262,144 tokens of context, 2× the 131,072 of Qwen3 235B A22B Thinking 2507.
- Nemotron 3 Ultra has 10 providers with a standard price, 7 more than Qwen3 235B A22B Thinking 2507; 2 providers sell both (Venice and GMI Cloud).
- On one account, Venice charges USD 1.25 for Nemotron 3 Ultra (1.37× its cheapest price) and USD 1.212 for Qwen3 235B A22B Thinking 2507 (its cheapest price) at 3:1, the lowest combined price among the providers that sell both.
- Shopping around matters more for Nemotron 3 Ultra: its dearest standard price is 1.37× its cheapest, against 1.08× for Qwen3 235B A22B Thinking 2507.
- Nemotron 3 Ultra has a batch price at 1 provider, from USD 0.4625 at 3:1 (Doubleword); Qwen3 235B A22B Thinking 2507 has no batch price on file.
- For the one-developer coding agent preset, the gap is USD 6,318 a year in Nemotron 3 Ultra's favour (Nemotron 3 Ultra USD 481.50 a month, Qwen3 235B A22B Thinking 2507 USD 1,008).
Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 prices side by side
USD per 1M tokens, standard tier. "Cheapest" is the lowest price at 3 input tokens to 1 output token; each line can come from a different provider.
| Price | Nemotron 3 Ultra | Qwen3 235B A22B Thinking 2507 |
|---|---|---|
| Model page | NVIDIA Nemotron 3 Ultra NVIDIA · 10 providers | Qwen3 235B A22B Thinking 2507 FP8 Qwen · 3 providers |
| Cheapest standard price, blended 3:1 | USD 0.9125 lower USD 0.50 in / USD 2.15 out at CoreWeave Forge CoreWeave Forge page checked 2026-11-15 | USD 1.212 USD 0.45 in / USD 3.50 out at Featherless and 1 other Featherless page checked 2026-11-15 |
| Cheapest input | USD 0.50 CoreWeave Forge source | USD 0.35 lower GMI Cloud source |
| Cheapest cached input | USD 0.10 CoreWeave Forge source | No standard price lists a cache price; cached tokens bill at the input price. |
| Cheapest output | USD 2.15 lower CoreWeave Forge source | USD 3.50 Featherless source |
| Maker's own API | NVIDIA publishes no paid standard price on its own API that we found. | Qwen publishes no paid standard price on its own API that we found. |
| Median standard price | USD 0.55 / USD 2.40 input / output, 10 prices | USD 0.45 / USD 3.50 input / output, 3 prices |
| Context window | 262,144 tokens, largest stated; from 256,000 at some providers | 131,072 tokens, largest stated; from 32,768 at some providers |
| Providers with a standard price | 10 | 3 |
- Lower marks the lower price on a line; tie, equal prices. Other tiers and free rows are on the model pages.
Report a wrong price or number (email to info@broadnet.ai, with this page's link filled in).
What do Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 cost for a real workload?
30 days at each model's cheapest standard price for the workload (presets of the cost calculator; cached input at the cached price where listed).
| Workload | Nemotron 3 Ultra | Qwen3 235B A22B Thinking 2507 | Difference |
|---|---|---|---|
| Support chatbot 2,000 requests a day, 1,500 input tokens (60% cached), 300 output | USD 62.10 CoreWeave Forge | USD 103.50 Featherless | USD 41.40 Nemotron 3 Ultra lower |
| RAG over documents 1,000 requests a day, 8,000 input tokens (30% cached), 500 output | USD 123.45 CoreWeave Forge | USD 147.00 GMI Cloud | USD 23.55 Nemotron 3 Ultra lower |
| Coding agent (one developer) 1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning | USD 481.50 CoreWeave Forge | USD 1,008 GMI Cloud | USD 526.50 Nemotron 3 Ultra lower |
Nemotron 3 Ultra is cheaper on all three workloads. The widest gap is on Coding agent (one developer): the dearer model costs 2.09× as much as the cheaper one.
Does the cheaper model change with your input:output ratio?
The cheaper model changes with the mix: below 14 input tokens per output token, Nemotron 3 Ultra is cheaper; above 14, Qwen3 235B A22B Thinking 2507 is cheaper. At the site's 3:1 blend, Nemotron 3 Ultra is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.
| Input : output tokens | Nemotron 3 Ultra | Qwen3 235B A22B Thinking 2507 |
|---|---|---|
| 1 : 3 (long answers) | USD 1.737 lower CoreWeave Forge | USD 2.738 Featherless |
| 1 : 1 | USD 1.325 lower CoreWeave Forge | USD 1.975 Featherless |
| 3 : 1 (the site’s blend) | USD 0.9125 lower CoreWeave Forge | USD 1.212 Featherless |
| 10 : 1 (retrieval, long prompts) | USD 0.65 lower CoreWeave Forge | USD 0.70 GMI Cloud |
| 30 : 1 (agents, big contexts) | USD 0.5532 CoreWeave Forge | USD 0.4742 lower GMI Cloud |
USD per 1M tokens at the cheapest standard provider for each mix, before cache discounts. Thinking tokens bill as output.
What do Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 cost per character in other languages?
Not measured: the real cost by language dataset covers only tokenizers we could measure from published files, and Qwen3 235B A22B Thinking 2507 is not among them. Count tokens on your own text with the counters below.
Check it on your own traffic
- NVIDIA Nemotron 3 Ultra API pricing: all 10 providers and other tiers.
- Qwen3 235B A22B Thinking 2507 FP8 API pricing: all 3 providers and other tiers.
- Cost calculator with Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 preloaded
- Nemotron 3 token counter
- Qwen token counter
- Prompt caching calculator
More comparisons: GLM 5.3 vs Nemotron 3 Ultra · Qwen3 235B A22B Thinking 2507 vs Qwen3 235B A22B 2507 · every Nemotron 3 Ultra comparison · every Qwen3 235B A22B Thinking 2507 comparison · all 662 price comparisons
Questions
Is Nemotron 3 Ultra cheaper than Qwen3 235B A22B Thinking 2507?
Yes. On a blend of 3 input tokens to 1 output token, Nemotron 3 Ultra costs USD 0.9125 per 1M tokens and Qwen3 235B A22B Thinking 2507 USD 1.212, at each model's cheapest standard price. For input alone, Qwen3 235B A22B Thinking 2507 is cheaper (USD 0.35 against USD 0.50 per 1M); for output alone, Nemotron 3 Ultra is (USD 2.15 against USD 3.50). This compares price only, not quality.
What do Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 cost per 1M tokens?
Nemotron 3 Ultra: from USD 0.50 input and USD 2.15 output at CoreWeave Forge; median of 10 standard prices USD 0.55 / USD 2.40. Qwen3 235B A22B Thinking 2507: from USD 0.45 input and USD 3.50 output at Featherless and 1 other; median of 3 standard prices USD 0.45 / USD 3.50. Each price is read from the provider's own pricing page; the source is linked in the table.
Which providers sell both Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507?
2 providers have a standard price for both: Venice (USD 1.25 and USD 1.212 at 3:1), GMI Cloud (USD 1.25 and USD 1.313 at 3:1). The first is the lowest combined price. Buying each model from its own cheapest provider costs USD 0.9125 and USD 1.212.
Which is cheaper for a coding agent or a chatbot, Nemotron 3 Ultra or Qwen3 235B A22B Thinking 2507?
For a coding agent (1,500 requests a day, 40,000 input tokens (85% cached), 2,000 output incl. reasoning): USD 481.50 a month on Nemotron 3 Ultra (CoreWeave Forge) against USD 1,008 on Qwen3 235B A22B Thinking 2507 (GMI Cloud), a difference of USD 526.50. For a support chatbot (2,000 requests a day, 1,500 input tokens (60% cached), 300 output): USD 62.10 a month on Nemotron 3 Ultra (CoreWeave Forge) against USD 103.50 on Qwen3 235B A22B Thinking 2507 (Featherless), a difference of USD 41.40. Each uses the cheapest standard price for that workload, cached input included. Try your own numbers.
At what input:output ratio does the cheaper of Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 change?
The cheaper model changes with the mix: below 14 input tokens per output token, Nemotron 3 Ultra is cheaper; above 14, Qwen3 235B A22B Thinking 2507 is cheaper. At the site's 3:1 blend, Nemotron 3 Ultra is cheaper. More input per output token, as with retrieval or agents that resend long contexts, favours the lower input price; long answers favour the lower output price.
Do Nemotron 3 Ultra and Qwen3 235B A22B Thinking 2507 have cached input pricing?
Nemotron 3 Ultra from USD 0.10 per 1M cached input tokens (CoreWeave Forge; 10 of 10 standard prices list a discount). Qwen3 235B A22B Thinking 2507: no standard price lists a cached-input discount. Cached tokens are billed at the input price where no cache price is published.
Which has the larger context window, Nemotron 3 Ultra or Qwen3 235B A22B Thinking 2507?
Nemotron 3 Ultra: up to 262,144 tokens, against 131,072 for Qwen3 235B A22B Thinking 2507, as stated on providers' pricing pages. Some providers offer less than the maximum; check the model pages before you pick the cheapest row for long prompts.