Pick Mercury 2.5 Preview for very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation or coding accuracy (95.7%, 91st percentile among cost-optimized models). Pick NVIDIA Nemotron 3 Ultra for the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48) or fast, efficient long-horizon agentic reasoning via a hybrid mamba-transformer design. Choose NVIDIA Nemotron 3 Ultra if you need self-hosting or data privacy; Mercury 2.5 Preview if you want a managed API.
Mercury 2.5 Preview (Inception Labs) and NVIDIA Nemotron 3 Ultra (NVIDIA) are two of the models people most often weigh against each other in 2026. Mercury 2.5 Preview is inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality. NVIDIA Nemotron 3 Ultra is nVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents. They diverge most on price, context window and open vs. closed weights — each quantified below from the models' real specs.
Key differences
Cost model: NVIDIA Nemotron 3 Ultra ships open weights you can self-host (hardware cost only, no per-token fee), while Mercury 2.5 Preview is API-metered at $0.04/$0.15 per 1M tokens. Your choice depends on whether you want zero marginal cost at the price of running infrastructure.
Context window: NVIDIA Nemotron 3 Ultra holds 3.8× more — 1M (~1,500 pages) vs 260K tokens (~390 pages). But effective recall usually fades long before the advertised ceiling, so the bigger number only helps if the model reasons over it.
Recency: Mercury 2.5 Preview is the newer model by about 3 months (released August 31, 2026), usually meaning fresher training data and capabilities.
Specifications
Spec
Mercury 2.5 Preview
NVIDIA Nemotron 3 Ultra
Provider
Inception Labs (US)
NVIDIA (US)
Released
August 31, 2026
June 4, 2026
Context window
260K tokens (~390 pages)
1M (~1,500 pages)
Price (in/out)
$0.04/$0.15 per 1M tokens
Open weight (self-host / free)
Open weight?
No — API only
Yes — self-hostable
Modalities
text
text, code
SWE-Bench Verified
Not published
Not published
MRCR v2 @ 1M
Not published
Not published
Who wins what
Very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation: Mercury 2.5 Preview — Inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality — and it is the newer of the two.
Coding accuracy (95.7%, 91st percentile among cost-optimized models): Mercury 2.5 Preview — NVIDIA Nemotron 3 Ultra is comparatively weak here — trails the best Chinese open models on overall intelligence
Mathematics accuracy (97.0%, 97th percentile): Mercury 2.5 Preview — Mercury 2.5 Preview lists mathematics accuracy (97.0%, 97th percentile) among its strengths; NVIDIA Nemotron 3 Ultra does not.
The most capable open-weight model from a US lab (Artificial Analysis Intelligence Index of about 48): NVIDIA Nemotron 3 Ultra — Open weights make this possible at all — Mercury 2.5 Preview is API-only, so it cannot leave the vendor's servers.
Fast, efficient long-horizon agentic reasoning via a hybrid Mamba-Transformer design: NVIDIA Nemotron 3 Ultra — Its 1M window holds about 3.8× more than Mercury 2.5 Preview's 260K tokens in a single prompt.
A fully open release — weights, training data, and recipes under a permissive license: NVIDIA Nemotron 3 Ultra — NVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents — and it carries the larger 1M context.
Lowest cost at scale: NVIDIA Nemotron 3 Ultra — Its weights are open, so at volume you pay for your own hardware instead of Mercury 2.5 Preview's $0.04/$0.15 per 1M tokens.
Largest single-prompt input: NVIDIA Nemotron 3 Ultra — Its 1M window is about 3.8× larger than Mercury 2.5 Preview's 260K tokens, fitting roughly 1,500 pages in one prompt.
Which should you pick?
A cost-sensitive startup shipping high volume: NVIDIA Nemotron 3 Ultra — At Open weight (self-host / free) it undercuts Mercury 2.5 Preview, and on millions of tokens that margin decides the monthly bill.
Someone analysing very long documents or codebases: NVIDIA Nemotron 3 Ultra — Larger 1M window fits more in one prompt.
A team with data-privacy or self-hosting needs: NVIDIA Nemotron 3 Ultra — Open weights let you run it on your own hardware; Mercury 2.5 Preview is API-only.
Anyone whose priority is very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation: Mercury 2.5 Preview — It is specifically built for that.
Anyone whose priority is the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48): NVIDIA Nemotron 3 Ultra — That is its strongest area.
Mercury 2.5 Preview: where it fits
Inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality. Released August 31, 2026 by Inception Labs, it is built for very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation, coding accuracy (95.7%, 91st percentile among cost-optimized models), mathematics accuracy (97.0%, 97th percentile), and tunable reasoning levels with parallel tool calls and schema-aligned JSON output.
Its trade-offs are real: 260K context window is far shorter than frontier 1M-token models, no vision or audio modalities, positioned only against cost-optimized models (GPT-5.6 Luna Low, Gemini 3.5 Flash-Lite), not frontier-class, and current $0.04/$0.15 pricing includes a limited-time launch promotion on OpenRouter (list price is $0.20/$0.75, may rise after Sept 8, 2026). At $0.04 in / $0.15 out per million tokens, it sits in the budget price band.
NVIDIA Nemotron 3 Ultra: where it fits
NVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents. Released June 4, 2026 by NVIDIA, it is built for the most capable open-weight model from a US lab (Artificial Analysis Intelligence Index of about 48), fast, efficient long-horizon agentic reasoning via a hybrid Mamba-Transformer design, a fully open release — weights, training data, and recipes under a permissive license, and strong coding for an open model (SWE-Bench Verified in the high 60s).
Its trade-offs: trails the best Chinese open models on overall intelligence, and a 550B mixture-of-experts is heavy to self-host, and the 1M context is rarely served in full. As an open-weight model, its running cost is your own hardware rather than a per-token fee.
The bottom line for this matchup
The defining split here is open vs. closed. NVIDIA Nemotron 3 Ultra gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Mercury 2.5 Preview gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.
Frequently asked questions
Is Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra better for coding?
Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Mercury 2.5 Preview leans toward very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation while NVIDIA Nemotron 3 Ultra leans toward the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48), and that positioning usually predicts which feels better on your codebase.
Which is cheaper, Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra?
NVIDIA Nemotron 3 Ultra is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Mercury 2.5 Preview is API-metered at $0.04/$0.15 per 1M tokens. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.
Which has the bigger context window?
NVIDIA Nemotron 3 Ultra — 1M vs 260K tokens, about 3.8× larger. Useful only if the model actually reasons over the full window, which not all do.
Can I use both Mercury 2.5 Preview and NVIDIA Nemotron 3 Ultra together?
Yes — a multi-model platform like LumiChats gives you Mercury 2.5 Preview, NVIDIA Nemotron 3 Ultra and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.
Which is newer, Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra?
Mercury 2.5 Preview — released August 31, 2026, about 3 months after NVIDIA Nemotron 3 Ultra.
Mercury 2.5 Preview vs NVIDIA Nemotron 3 Ultra
Inception Labs · US | NVIDIA · US · Updated June 2026
Quick verdict
Pick Mercury 2.5 Preview for very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation or coding accuracy (95.7%, 91st percentile among cost-optimized models). Pick NVIDIA Nemotron 3 Ultra for the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48) or fast, efficient long-horizon agentic reasoning via a hybrid mamba-transformer design. Choose NVIDIA Nemotron 3 Ultra if you need self-hosting or data privacy; Mercury 2.5 Preview if you want a managed API.
Mercury 2.5 Preview (Inception Labs) and NVIDIA Nemotron 3 Ultra (NVIDIA) are two of the models people most often weigh against each other in 2026. Mercury 2.5 Preview is inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality. NVIDIA Nemotron 3 Ultra is nVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents. They diverge most on price, context window and open vs. closed weights — each quantified below from the models' real specs.
Key differences at a glance
▸Cost model: NVIDIA Nemotron 3 Ultra ships open weights you can self-host (hardware cost only, no per-token fee), while Mercury 2.5 Preview is API-metered at $0.04/$0.15 per 1M tokens. Your choice depends on whether you want zero marginal cost at the price of running infrastructure.
▸Context window: NVIDIA Nemotron 3 Ultra holds 3.8× more — 1M (~1,500 pages) vs 260K tokens (~390 pages). But effective recall usually fades long before the advertised ceiling, so the bigger number only helps if the model reasons over it.
▸Recency: Mercury 2.5 Preview is the newer model by about 3 months (released August 31, 2026), usually meaning fresher training data and capabilities.
Side-by-side specs
Spec
Mercury 2.5 Preview
NVIDIA Nemotron 3 Ultra
Provider
Inception Labs (US)
NVIDIA (US)
Released
August 31, 2026
June 4, 2026
Context window
260K tokens (~390 pages)
1M (~1,500 pages)
Price (in/out)
$0.04/$0.15 per 1M tokens
Open weight (self-host / free)
Open weight?
No — API only
Yes — self-hostable
Modalities
text
text, code
SWE-Bench Verified
Not published
Not published
MRCR v2 @ 1M
Not published
Not published
Who wins what
Very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation
Mercury 2.5 Preview
Inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality — and it is the newer of the two.
Coding accuracy (95.7%, 91st percentile among cost-optimized models)
Mercury 2.5 Preview
NVIDIA Nemotron 3 Ultra is comparatively weak here — trails the best Chinese open models on overall intelligence
Mathematics accuracy (97.0%, 97th percentile)
Mercury 2.5 Preview
Mercury 2.5 Preview lists mathematics accuracy (97.0%, 97th percentile) among its strengths; NVIDIA Nemotron 3 Ultra does not.
The most capable open-weight model from a US lab (Artificial Analysis Intelligence Index of about 48)
NVIDIA Nemotron 3 Ultra
Open weights make this possible at all — Mercury 2.5 Preview is API-only, so it cannot leave the vendor's servers.
Fast, efficient long-horizon agentic reasoning via a hybrid Mamba-Transformer design
NVIDIA Nemotron 3 Ultra
Its 1M window holds about 3.8× more than Mercury 2.5 Preview's 260K tokens in a single prompt.
A fully open release — weights, training data, and recipes under a permissive license
NVIDIA Nemotron 3 Ultra
NVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents — and it carries the larger 1M context.
Lowest cost at scale
NVIDIA Nemotron 3 Ultra
Its weights are open, so at volume you pay for your own hardware instead of Mercury 2.5 Preview's $0.04/$0.15 per 1M tokens.
Largest single-prompt input
NVIDIA Nemotron 3 Ultra
Its 1M window is about 3.8× larger than Mercury 2.5 Preview's 260K tokens, fitting roughly 1,500 pages in one prompt.
Which should you pick?
A cost-sensitive startup shipping high volume
→ NVIDIA Nemotron 3 Ultra
At Open weight (self-host / free) it undercuts Mercury 2.5 Preview, and on millions of tokens that margin decides the monthly bill.
Someone analysing very long documents or codebases
→ NVIDIA Nemotron 3 Ultra
Larger 1M window fits more in one prompt.
A team with data-privacy or self-hosting needs
→ NVIDIA Nemotron 3 Ultra
Open weights let you run it on your own hardware; Mercury 2.5 Preview is API-only.
Anyone whose priority is very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation
→ Mercury 2.5 Preview
It is specifically built for that.
Anyone whose priority is the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48)
→ NVIDIA Nemotron 3 Ultra
That is its strongest area.
Mercury 2.5 Preview: where it fits
Inception Labs' August 31, 2026 diffusion-based language model preview, refining tokens in parallel for roughly 10x the throughput of comparable autoregressive models at cost-optimized-tier quality. Released August 31, 2026 by Inception Labs, it is built for very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation, coding accuracy (95.7%, 91st percentile among cost-optimized models), mathematics accuracy (97.0%, 97th percentile), and tunable reasoning levels with parallel tool calls and schema-aligned JSON output.
Its trade-offs are real: 260K context window is far shorter than frontier 1M-token models, no vision or audio modalities, positioned only against cost-optimized models (GPT-5.6 Luna Low, Gemini 3.5 Flash-Lite), not frontier-class, and current $0.04/$0.15 pricing includes a limited-time launch promotion on OpenRouter (list price is $0.20/$0.75, may rise after Sept 8, 2026). At $0.04 in / $0.15 out per million tokens, it sits in the budget price band.
NVIDIA Nemotron 3 Ultra: where it fits
NVIDIA's open-weight reasoning flagship (about 550B total, 55B active) — the most capable open model from a US lab, built for long-running agents. Released June 4, 2026 by NVIDIA, it is built for the most capable open-weight model from a US lab (Artificial Analysis Intelligence Index of about 48), fast, efficient long-horizon agentic reasoning via a hybrid Mamba-Transformer design, a fully open release — weights, training data, and recipes under a permissive license, and strong coding for an open model (SWE-Bench Verified in the high 60s).
Its trade-offs: trails the best Chinese open models on overall intelligence, and a 550B mixture-of-experts is heavy to self-host, and the 1M context is rarely served in full. As an open-weight model, its running cost is your own hardware rather than a per-token fee.
The bottom line for this matchup
The defining split here is open vs. closed. NVIDIA Nemotron 3 Ultra gives you weights you control — self-host it, fine-tune it, keep data in-house, pay only for hardware. Mercury 2.5 Preview gives you a managed, always-updated API with no infrastructure to run. Teams with GPUs, privacy requirements, or huge volume often favour the open model; teams that want zero ops and the latest capabilities favour the closed one. Capability is close enough that this operational question, not the benchmark, usually decides it.
Want both Mercury 2.5 Preview and NVIDIA Nemotron 3 Ultra without two subscriptions? LumiChats gives you these plus 40+ models under one ₹69/day pass (about $1/day) — draft with one, cross-check with the other.
Is Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra better for coding?
Public SWE-Bench figures are not available for either model, so the honest test is your own repository — run an identical real bug through both. By design, Mercury 2.5 Preview leans toward very high inference throughput (1,107 tok/s) via diffusion-based parallel token generation while NVIDIA Nemotron 3 Ultra leans toward the most capable open-weight model from a us lab (artificial analysis intelligence index of about 48), and that positioning usually predicts which feels better on your codebase.
Which is cheaper, Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra?
NVIDIA Nemotron 3 Ultra is open-weight, so self-hosting means no per-token fee (you pay for hardware instead), while Mercury 2.5 Preview is API-metered at $0.04/$0.15 per 1M tokens. For most teams without GPUs, the API model is cheaper to start; at very high volume, self-hosting can win.
Which has the bigger context window?
NVIDIA Nemotron 3 Ultra — 1M vs 260K tokens, about 3.8× larger. Useful only if the model actually reasons over the full window, which not all do.
Can I use both Mercury 2.5 Preview and NVIDIA Nemotron 3 Ultra together?
Yes — a multi-model platform like LumiChats gives you Mercury 2.5 Preview, NVIDIA Nemotron 3 Ultra and 40+ others under one ₹69/day pass (about $1/day), so you can draft with one and cross-check with the other instead of buying two subscriptions.
Which is newer, Mercury 2.5 Preview or NVIDIA Nemotron 3 Ultra?
Mercury 2.5 Preview — released August 31, 2026, about 3 months after NVIDIA Nemotron 3 Ultra.
Specifications and benchmarks reflect publicly reported figures as of June 2026 and may change as providers release updates. Always verify on your own workload.