Model comparison

Kimi K3 vs NVIDIA Nemotron 3 Ultra

Moonshot AI against NVIDIA, compared on context, price, and verified benchmark results.

Catalog record checked September 3, 2026Individual provider fields may changeEvidence confidence: High — see receipts below

Moonshot AI

Kimi K3

Frontier · Open weights

NVIDIA

NVIDIA Nemotron 3 Ultra

Frontier · Open weights

AI model capability comparison
Specification	Kimi K3	NVIDIA Nemotron 3 Ultra
Provider	Moonshot AI	NVIDIA
Tier	Frontier	Frontier
Context window	Winner: 1.05M	1M
Max output	128K	Not verifiedUnverified
Input / 1M tokens	$3	Winner: $0.50
Output / 1M tokens	$15	Winner: $2.20
Weights	Open	Open
Parameters	2.8T (open-weight MoE)	550B total / 55B active (MoE)
Reasoning levels	low, high, max	low, high, max
Modalities	text, image, video	text
API model id	kimi-k3	nemotron-3-ultra-550b-a55b
Released	July 16, 2026	June 4, 2026
Artificial Analysis Intelligence Index [max] (2026-08-14)	60	Not verifiedUnverified
Artificial Analysis Intelligence Index (2026-08-14)	Not verifiedUnverified	54
Terminal-Bench 2.1 (2026-07-27)	88.3	Not verifiedUnverified
GPQA Diamond (2026-07-27)	93.5	Not verifiedUnverified
Humanity's Last Exam (2026-07-27)	56	Not verifiedUnverified
SWE-bench Verified (2026-09-01)	93.4	Not verifiedUnverified

Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Bold indicates the better figure where one is objectively better. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.

Benchmark receipts

2026-09-01: vals.ai SWE-bench Verified leaderboard (independent, mini-swe-agent bash-only); not published by Moonshot; a 76.8 aggregator claim is low-reliability — both notedReal GitHub issue resolution: does the model's patch pass the hidden tests. Comparability: comparable with caveat — Post-audit vendor claims and pre-audit scores sit on different task trust levels; scaffolding (agent harness, compute budget) also dominates results. Never aggregate across scaffolds.
2026-08-14: Artificial AnalysisComposite index blending reasoning, knowledge, and coding evals into one 0–100 score. Comparability: directly comparable — AA occasionally rebaselines the index scale between snapshots — a score captured on one date is only comparable to same-snapshot scores (check measuredAt).
2026-07-27: Kimi K3 HF model card (vendor, with tools; 43.5 no tools)Agentic terminal work: multi-step tasks executed in a sandboxed shell environment. Comparability: comparable with caveat — Scores across Terminal-Bench major versions (v1 vs v2) are NOT comparable; harness configuration (container, time limits) also shifts results.

Pricing tiers: Kimi K3: $3/$15 per MTok; cached input $0.30/MTok. Open-weight under the custom Kimi K3 License (not a standard open-source licence). 2.8T MoE with native vision. · NVIDIA Nemotron 3 Ultra: Open weights (weights + data + recipes); $0.50/$2.20 per MTok is the third-party hosted rate (OpenRouter). 550B/55B MoE.

FrontierOpen weightsRecord checked September 3, 2026

Kimi K3

Kimi K3 is Moonshot AI's open-weight frontier model — 2.8T MoE, 1M context, tied near the top of open-weight leaderboards.

Best for

Open-weight frontier work
Long-context
Multimodal

Watch out

Self-hosting needs large clusters; hosted rates vary.

FrontierOpen weightsRecord checked September 2, 2026

NVIDIA Nemotron 3 Ultra

Nemotron 3 Ultra is NVIDIA's open-weight (550B/55B) frontier MoE with a 1M-token context.

Best for

Open-weight frontier
Self-hosting on NVIDIA GPUs
Reasoning

Watch out

Top US open-weight until Thinking Machines Inkling; verify hosted pricing.

When the cheaper one wins

NVIDIA Nemotron 3 Ultra is cheaper on output at $2.20 per million tokens against $15 for Kimi K3 — about 6.8×. Use the cheaper tier for classification, extraction, summarisation, and any task where the expensive model’s extra score does not change the accepted output. The expensive one only pays if your hardest task actually fails on the cheap tier. These are standard-tier API rates, excluding batch and cache discounts.

Run the model picker

Evidence confidence: High

How strong and complete the evidence behind this comparison is — not a prediction of which model is better.

Pricing verified on both sides — Input and output rates are verified for both models.
4/5 core specs verified on both sides — Not published for at least one side: max output.
1 shared named benchmark with differing scores — Measured on: Artificial Analysis Intelligence Index.
Verified within the last 90 days — Newest catalog check was 8 days ago.
Both models carry source citations — Each side has at least two catalog sources on record.

Source receipts

Each catalog figure was checked against the provider or an independent second source on the date shown.

Kimi K3: HuggingFace — Kimi K3 (accessed 2026-08-29)
Kimi K3: Moonshot — Kimi K3 guide (accessed 2026-08-29)
NVIDIA Nemotron 3 Ultra: NVIDIA — Nemotron (accessed 2026-08-29)
NVIDIA Nemotron 3 Ultra: NVIDIA Nemotron 3 Ultra (accessed 2026-08-29)

Related comparisons

Diving deeper on one model? Kimi K3 · NVIDIA Nemotron 3 Ultra

Common questions

Kimi K3 vs NVIDIA Nemotron 3 Ultra

Answered from the verified figures on this page rather than general guidance.

Is Kimi K3 or NVIDIA Nemotron 3 Ultra cheaper for input?

NVIDIA Nemotron 3 Ultra is cheaper at $0.50 per million input tokens, against $3 for Kimi K3 — roughly 6.0× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Kimi K3 has tiered pricing: $3/$15 per MTok; cached input $0.30/MTok. Open-weight under the custom Kimi K3 License (not a standard open-source licence). 2.8T MoE with native vision. NVIDIA Nemotron 3 Ultra has tiered pricing: Open weights (weights + data + recipes); $0.50/$2.20 per MTok is the third-party hosted rate (OpenRouter). 550B/55B MoE.

Is Kimi K3 or NVIDIA Nemotron 3 Ultra cheaper for output?

NVIDIA Nemotron 3 Ultra is cheaper at $2.20 per million output tokens, against $15 for Kimi K3 — roughly 6.8× the price. Output tokens usually dominate a real bill, so weigh the output rate more heavily than the input rate. These are base rates; Kimi K3 has tiered pricing: $3/$15 per MTok; cached input $0.30/MTok. Open-weight under the custom Kimi K3 License (not a standard open-source licence). 2.8T MoE with native vision. NVIDIA Nemotron 3 Ultra has tiered pricing: Open weights (weights + data + recipes); $0.50/$2.20 per MTok is the third-party hosted rate (OpenRouter). 550B/55B MoE.

Which has the larger context window, Kimi K3 or NVIDIA Nemotron 3 Ultra?

Kimi K3 accepts 1.05M tokens against 1M for NVIDIA Nemotron 3 Ultra. This only matters if you routinely send very long documents or large codebases.

Do Kimi K3 and NVIDIA Nemotron 3 Ultra support the same reasoning levels?

Yes — both accept the same effort settings: "low", "high", "max". Higher effort costs more and takes longer, so start low and raise it only where output quality actually improves.

Should I use Kimi K3 or NVIDIA Nemotron 3 Ultra?

Both sit in the frontier tier, so the choice usually comes down to price and context rather than capability. Kimi K3 suits open-weight frontier work; NVIDIA Nemotron 3 Ultra suits open-weight frontier.

Next step

Choosing between them

The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.

Input price: NVIDIA Nemotron 3 Ultra costs $0.50 per 1M tokens versus $3 for Kimi K3 — a 6x difference at the headline tier.
Context: Kimi K3 takes 1.05M against 1M for NVIDIA Nemotron 3 Ultra — only decisive if your prompts approach the smaller window.
Measured capability: Kimi K3 leads Artificial Analysis Intelligence Index 60 to 54 (measured 2026-08-14).

Run the model picker Estimate stack cost All model comparisons

Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.