Model comparison
Claude Fable 5.1 vs NVIDIA Nemotron 3 Ultra
Anthropic against NVIDIA, compared on context, price, and attributed benchmark results.
Claude Fable 5.1: profile checked October 25, 2026NVIDIA Nemotron 3 Ultra: profile checked October 25, 2026Individual provider fields may changeEvidence confidence: Insufficient data — see receipts below
Anthropic
Claude Fable 5.1
Frontier
NVIDIA
NVIDIA Nemotron 3 Ultra
Frontier · Open weights
| Specification | Claude Fable 5.1 | NVIDIA Nemotron 3 Ultra |
|---|---|---|
| Provider | ||
| Provider | Anthropic | NVIDIA |
| Tier | ||
| Tier | Frontier | Frontier |
| Context window | ||
| Context window | 1M | Winner: 1.05M |
| Max output | ||
| Max output | 128K | Not verifiedUnverified |
| Input / 1M tokens | ||
| Input / 1M tokens | $10 | Not verifiedUnverified |
| Output / 1M tokens | ||
| Output / 1M tokens | $50 | Not verifiedUnverified |
| Weights | ||
| Weights | Closed | Open |
| Parameters | ||
| Parameters | Not verifiedUnverified | 550B total / 55B active (MoE) |
| Reasoning levels | ||
| Reasoning levels | low, medium, high, xhigh, max | Not verifiedUnverified |
| Modalities | ||
| Modalities | text, image | text |
| License | ||
| License | Not verifiedUnverified | OpenMDW License Agreement 1.1 |
| API model id | ||
| API model id | claude-fable-5-1 | nvidia/nemotron-3-ultra-550b-a55b |
| Released | ||
| Released | September 1, 2026 | June 4, 2026 |
| Artificial Analysis Intelligence Index [max] · source date Claude Fable 5.1: 2026-10-01 | ||
| Artificial Analysis Intelligence Index [max] · source date Claude Fable 5.1: 2026-10-01 | 53 | Not verifiedUnverified |
| Artificial Analysis Intelligence Index · source date NVIDIA Nemotron 3 Ultra: 2026-10-01 | ||
| Artificial Analysis Intelligence Index · source date NVIDIA Nemotron 3 Ultra: 2026-10-01 | Not verifiedUnverified | 23 |
| Humanity's Last Exam · source date Claude Fable 5.1: 2026-09-01 | ||
| Humanity's Last Exam · source date Claude Fable 5.1: 2026-09-01 | 65 | Not verifiedUnverified |
| Terminal-Bench 2.1 · source date Claude Fable 5.1: 2026-09-26 | ||
| Terminal-Bench 2.1 · source date Claude Fable 5.1: 2026-09-26 | 91.4 | Not verifiedUnverified |
| DeepSWE 1.1 · source date Claude Fable 5.1: 2026-09-03 | ||
| DeepSWE 1.1 · source date Claude Fable 5.1: 2026-09-03 | 67.4 | Not verifiedUnverified |
Dates beside benchmark scores refer to the linked source’s publication or capture date unless that source explicitly states when it measured the result. A shared URL alone does not establish comparable runs; unverified setups are marked. Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.
Benchmark receipts
Where each score comes from, and how far it can be compared across models.
- Source date2026-10-01Artificial Analysis (Nemotron 3 Ultra)
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
- Source date2026-10-01Artificial Analysis (Claude Fable 5.1 (max with fallback))
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
- Source date2026-09-26Artificial Analysis Terminal-Bench 2.1 (independent, Terminus 2, max effort)
Agentic terminal work: multi-step tasks executed in a sandboxed shell environment.
Comparable with caveatNot comparable with Terminal-Bench 3.0 or 4.0 (different task sets) or v1; harness configuration (container, time limits) also shifts results.
- Source date2026-09-03OpenAI GPT-6 Astra launch table quoting a cross-provider result; not a direct Datacurve row (Cross-provider quotation; Fable harness, effort and experiment date unverified)
Agentic software engineering: pass@1 on real repo tasks with reported cost per completed task.
Directly comparable
- Source date2026-09-01Anthropic Fable 5.1 announcement (vendor, with tools; 60.9 no tools) — Anthropic's later Opus 5.5 table lists 65.6; conflict preserved
Frontier-knowledge ceiling: extremely hard multi-domain questions written to be near-impossible without deep expertise.
Comparable with caveatSubset/tool configurations (text-only vs with-tools) differ between vendors and materially change scores; must match configuration to compare.
Pricing tiers
Claude Fable 5.1: Standard USD/MTok input/output $10/$50 with no long-context surcharge; cache reads $0.25 versus Fable 5's $1, 5-minute writes $12.50, 1-hour writes $20. Batch $5/$25. US-only inference adds 10% to input/output and cache reads/writes on Anthropic-operated routes where supported; Foundry uses US Data Zone Standard. Bedrock and Google-managed Google Cloud have independent regional prices. Adaptive thinking is always on; default effort high. Knowledge cutoff June 2026; retirement not sooner than September 1, 2027. Server-side tools can add usage fees; tool schemas/results and thinking consume billable tokens.
NVIDIA Nemotron 3 Ultra: BF16 and NVFP4 weights use OpenMDW 1.1; self-hosting incurs compute/storage costs. No current NVIDIA-direct USD/token tariff or universal free quota is established. The published BF16 recipe starts at 262,144 tokens and supports 1,048,576 with explicit vLLM/SGLang long-context settings; this is a documented serving extension, not a freshly verified NVFP4 config default. The NVFP4 config exceeded the existing 5 MiB capture cap. Tinker’s separately priced 64K/256K BF16 training rows, including its limited-time discount, do not establish NVIDIA-direct inference prices or native context limits.
Claude Fable 5.1
Claude Fable 5.1 is Anthropic's model for the most demanding reasoning and long-horizon agentic work at $10/$50 — Anthropic now recommends Opus 5.5 as the default for most workloads.
Best for
- Frontier agents
- Hardest reasoning work
- Fable 5 workloads at the same price
Watch out
Standard token rates are 2.5x Opus 5.5's. Forced tool use errors; thinking blocks are model/conversation-bound, so editing earlier turns or switching models can invalidate them. Display updates and per-message effort are beta. Match benchmark settings and evaluate pinned workloads before migrating.
NVIDIA Nemotron 3 Ultra
Nemotron 3 Ultra is NVIDIA's open-weight 550B/55B hybrid Mamba-attention MoE with BF16/NVFP4 checkpoints and publisher-documented serving extension up to 1,048,576 tokens.
Best for
- Open-weight frontier
- Self-hosting on NVIDIA GPUs
- Reasoning
Watch out
The current BF16 card documents enable_thinking on/off, enabled by default, and a separate medium_effort example control. The NIM example generator maps effort choices into those kwargs; a universal native graded-effort enum is not established. The card’s reasoning_budget wrapper closes at a newline or up to 500 tokens past its budget; it is not a universal hosted parameter or hard output cap. Keep BF16/NVFP4 hardware and evaluation configurations distinct. NIM sample max_tokens values are request examples, not verified ceilings. Redistribution under OpenMDW 1.1 retains the agreement and applicable origin notices; the agreement imposes no restrictions on use/modification/sharing of outputs. Hosted service terms are separate.
When the cheaper one wins
One of these rows is missing a verified output price, so this page will not name a cheaper winner. Run the model picker against the actual job instead of guessing.
Run the model pickerEvidence confidence: Insufficient data
How strong and complete the evidence behind this comparison is — not a prediction of which model is better. Hard limitations apply: missing pricing; evidence too sparse.
- Pricing incomplete — No verified token pricing for NVIDIA Nemotron 3 Ultra.
- 2/5 core specs verified on both sides — Not published for at least one side: max output, parameter count, reasoning levels.
- No directly aligned benchmark setup — No benchmark has matching cited sources, source dates, units, effort and recorded run details on both models.
- Verified within the last 90 days — Older of the two catalog checks was 24 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.
- Claude Fable 5.1: Whole-record source check: exact model card (accessed 2026-10-25)
- Claude Fable 5.1: Whole-record source check: current publisher pricing, controls or access terms (accessed 2026-10-25)
- Claude Fable 5.1: Whole-record source check: current publisher pricing, controls or access terms (accessed 2026-10-25)
- Claude Fable 5.1: Whole-record source check: current publisher pricing, controls or access terms (accessed 2026-10-25)
- Claude Fable 5.1: Anthropic models overview — Fable 5.1 (1M context, 128K output) (accessed 2026-09-03)
- Claude Fable 5.1: Anthropic — Claude Fable and Mythos 5.1 announcement (accessed 2026-09-26)
- Claude Fable 5.1: Anthropic models overview — Opus 5.5 default; Fable 5.1 for demanding reasoning (accessed 2026-09-26)
- NVIDIA Nemotron 3 Ultra: Current exact native checkpoint, documented serving recipe, hosted schema or governing licence (accessed 2026-10-25)
- NVIDIA Nemotron 3 Ultra: Current exact native checkpoint, documented serving recipe, hosted schema or governing licence (accessed 2026-10-25)
- NVIDIA Nemotron 3 Ultra: Current exact native checkpoint, documented serving recipe, hosted schema or governing licence (accessed 2026-10-25)
- NVIDIA Nemotron 3 Ultra: Current exact native checkpoint, documented serving recipe, hosted schema or governing licence (accessed 2026-10-25)
- NVIDIA Nemotron 3 Ultra: Current exact native checkpoint, documented serving recipe, hosted schema or governing licence (accessed 2026-10-25)
- NVIDIA Nemotron 3 Ultra: NVIDIA exact NVFP4 card: release, serving limits, thinking and vendor benchmarks (accessed 2026-10-02)
- NVIDIA Nemotron 3 Ultra: NVIDIA raw checkpoint config: 262,144 default positions (accessed 2026-10-02)
Related comparisons
- Claude Fable 5.1 vs Claude Haiku 4.5
- Claude Fable 5.1 vs Claude Haiku 5.5
- Claude Fable 5.1 vs Claude Mythos 5.1
- Claude Fable 5.1 vs Claude Opus 5
- Claude Fable 5.1 vs Claude Opus 5.5
- Claude Fable 5.1 vs Claude Sonnet 5
Diving deeper on one model? Claude Fable 5.1 · NVIDIA Nemotron 3 Ultra
Common questions
Claude Fable 5.1 vs NVIDIA Nemotron 3 Ultra
Answered from the verified figures on this page rather than general guidance.
Which has the larger context window, Claude Fable 5.1 or NVIDIA Nemotron 3 Ultra?
Should I use Claude Fable 5.1 or NVIDIA Nemotron 3 Ultra?
Can I self-host Claude Fable 5.1 or NVIDIA Nemotron 3 Ultra?
Next step
Choosing between them
The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.
- Context: NVIDIA Nemotron 3 Ultra takes 1.05M against 1M for Claude Fable 5.1 — only decisive if your prompts approach the smaller window.
- Deployment: NVIDIA Nemotron 3 Ultra publishes weights you can self-host; the other is API-only.
Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.