How does R1 rank among AI models?
R1 is a superseded version and is not on the current board as of 6 November 2026. Its archive position, where one exists, is shown on the page; the newer version carries the current rank.
DeepSeek·#200 of 504 in the all-versions archive
R1 is a language model from DeepSeek. It ranks #200 of 504 in ModelCap's all-versions archive, with an Index score of 51.7, measured on 3 public benchmarks. Its score range of 46.8–56.6 spans positions #165–#233. It reads up to 64K tokens of context, and its weights are openly downloadable.
Figures as of 6 Nov 2026, 20:49 UTC · download the public dataset · methodology
ModelCap Index · All-versions archive
51.7#200
Measured on 3 public benchmarks · range 46.8–56.6
Evidence as of 6 Nov 2026, 19:11 UTC
Superseded by R1 0528. It no longer holds a current-board position; its all-versions archive rank is #200.
Input
Self-hosted
weights only · no listed API price
Output
Self-hosted
weights only · no listed API price
Context
64K
tokens
Providers
0
endpoints
Providers
Uptime measured over the last 30 minutesNo provider is currently serving this model.
Hugging Face Inference Providers
Exact repository identityProvider economics and telemetry published by the Hugging Face Router. These are operational signals, not benchmark results.
Reported evaluation leads (4)
Not independent measurementStructured rows reported in this exact model repository at the pinned revision below. They are the publisher talking about itself: never measured evidence, never a score input on their own. They remain quarantined until dataset revision, harness, configuration and identity pass ModelCap’s independent admission review.
Other versions (1)
ModelCap Index · All-versions archive
Methodology51.7
Measured on 3 public benchmarks
All-versions archive
#200 of 504
- Score range
- 46.8–56.6
- Rank range
- #165–#233
- Evidence as of
- 6 Nov 2026, 19:11 UTC
- Benchmarks used
- ARC-AGI-2 Verified, Arena Coding, Arena Overall
Technical details
- Public rank basis
- Measured (measured)
- Evidence label
- Measured · 3 public benchmark observations across 3 boards
- Index support
- 63.9% · strong
- Identity binding
- Exact catalogue product · deepseek/deepseek/deepseek-r1/fp8 · aggregates disclosed benchmark configurations · endpoint configuration not claimed · artifact metadata deepseek-ai/DeepSeek-R1@56d4cbbb4d29f4355bab4b9a39ccb717a14ad5ad (not the evaluation revision)
- Index observations used
- 3
- Qualified Index evidence point
- 51.7
Bayesian Evidence Score · measured evidence input only
BES normalizes admitted public benchmark evidence for measured Index rows. Its legacy score and observables prior do not define the public language rank.
- Observed capability
- 51.7
- Evidence status
- confirmed · 86% mass
- General preference
- 52.5
- Coding
- 52.0
- Agents & tools
- —
- Reasoning
- 13.7
- Evidence breadth
- 95%
- Evidence coverage
- 64%
Market and catalogue signals · secondary, never the language Index rank
- Usage (OpenRouter popularity)
- —
- Liquidity (providers × uptime)
- —
- Open reach (HF downloads)
- 98.0
- Surface (context / tools / modalities)
- 65.0
- Freshness
- 8.2
- · HF 30d downloads 1641567
- · Catalogue: context 64000, reasoning, tools
- · First seen 2025-01-20
These adoption and deployment observations remain context only. They do not change this model's ModelCap Index score or rank.
BES evidence selected · measured input only
- Arena Overall52.5
- Arena Coding52.0
- ARC-AGI-2 Verified13.7
The bounded 0–100 capability score blends each source’s competitive placement with its published achievement, then combines capability families. Evidence breadth adds a modest uncertainty adjustment; price and popularity are not part of benchmark-led rank.
Deployment readiness
Metadata index78.6
strong · 88% metadata coverage
- Artifact reproducibility
- 85.0
- Access & legal clarity
- 100.0
- Deployability
- 60.0
- Evaluation provenance
- 75.4
Not yet published: the base model it derives from, measured provider uptime, pinned evaluation dataset versions.
A metadata completeness and deployability index—not a safety certification, quality grade, or production-readiness claim.
Market activity
Venue coverage- Market Gravity
- 15.1
- OpenRouter weekly popularity
- Not listed
Market Gravity uses OpenRouter’s full-catalogue popularity order. Sparse Vercel top-ten observations appear only when published and do not affect the score.
Arena preference
- Status
- Ranked
- Source
- LMArena
- Official rank
- #168 of 419
- Rating
- 1398.1
- 95% confidence interval
- 1393–1403
- Votes
- 19K
- Source published at
- 28 Oct 2026, 00:00 UTC
- Category
- overall
- Identity match confidence
- 99%
Market Gravity breakdown
- Usage55%
- 0.0
- Liquidity25%
- 0.0
- Open reach15%
- 98.0
- Freshness5%
- 8.2
OpenRouter weekly popularity across the full model catalogue.
Independent providers versus the model's open or closed cohort, weighted by uptime.
Hugging Face 30-day downloads for open models; neutral for closed models.
Time since first public availability, on a six-month half-life.
Specification
- Model ID
deepseek/deepseek-r1- Context window
- 64K tokens
- Declared output cap
- 16K tokens
OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. OpenRouter metadata
- Inputs
- Text
- Outputs
- Text
- Tokenizer
- DeepSeek
- Cached input tokens
- Not offered
- First seen on OpenRouter
- 20 Jan 2025
- Knowledge cutoff
- 2024-07-31
Capabilities
- Supported: Reasoning
- Supported: Tool use
- Not supported: Structured output
- Supported: Response format
- Not supported: Moderated
Weights & access
Open weights. Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI.
- Access
- Open weights
- Licence
- mit
- Pinned revision
- 56d4cbbb4d29
- Parameters
- 684.5B
- Architecture
- DeepseekV3ForCausalLM
- Model type
- deepseek_v3
- Hugging Face downloads (30d)
- 1.6M
- Downloads (all time)
- 51.2M
- Likes
- 14K
- Repository
- deepseek-ai/DeepSeek-R1
- Repository updated
- 27 Mar 2025, 04:01 UTC
Benchmark leaderboards featuring R1
All leaderboardsR1: common questions
How much does R1 cost per 1M tokens?
No live API price is listed for R1 in the current snapshot as of 6 November 2026 because no serving endpoint is published. ModelCap shows a price only when a provider currently lists one.
What is the context window of R1?
R1 has a published context window of 64K tokens (64,000). The snapshot lists a 16,000 token catalogue-declared route output cap. OpenRouter catalogue or provider-endpoint declaration. Limits can vary by provider and prompt length; a model-wide maximum has not been independently verified. The providers table lists each endpoint's context limit.
Is R1 open-weight?
Open weights (mit). Downloadable weights are published under an OSI or free-culture license in the current repository metadata. This describes weight availability, not whether the full training stack qualifies as Open Source AI. The repository is deepseek-ai/DeepSeek-R1 on Hugging Face.
Which benchmarks has R1 been evaluated on?
R1 has published results on 3 public boards tracked by ModelCap as of 6 November 2026: Arena text 1398 (#168 of 419), Arena coding 1444 (#167 of 414) and ARC-AGI-2 1.3% (#234 of 281). Each board page ranks every tracked model on that benchmark. The ModelCap Index scores the results its methodology admits, with published uncertainty, and shows reference boards without scoring them.
When was R1 released?
R1 first appeared in the catalogue on 20 Jan 2025, with a published knowledge cutoff of 2024-07-31. It has since been superseded by R1 0528. Rank and price movements since then are recorded on the site's changes feed.