Model
DeepSeek-R1
Classic open reasoning model for transparent math, coding, and deliberate problem solving
Compare providers (sorted by listed input price)
15 of 19 providers publish a price for DeepSeek-R1. The cheapest full-context option is SiliconFlow (China) at $0.5/$2.18 per M. NanoGPT lists lower but caps context at 128K.
Best price history
Daily lowest published price across live providers. A lower line is better for the buyer.
- Best input /M
- $0.4unchanged
- Best output /M
- $1.70unchanged
- Provider coverage
- 21
- 21 at the start · 24 snapshots
Published best prices are unchanged.
24 daily snapshots collected from Aug 21, 2026 through Sep 14, 2026. A price chart will appear after the first real move.
Price badge
https://www.vaanalytics.in/badge/deepseek/deepseek-r1.svgLive SVG, regenerated on every hourly sync — paste it into a README to always show the current cheapest listed price. Updates when prices change; no build step needed on your side.
Recent changes
Changelog →- capabilityvia amazon-bedrock · Sep 13, 2026
- tool_call: true → false
- capabilityvia vercel · Sep 13, 2026
- description: DeepSeek reasoning model for multi-step analysis, math, coding, and tools → Classic open reasoning model for transparent math, coding, and deliberate problem solving
- tool_call: true → false
- open_weights: false → true
- capabilityvia openrouter · Sep 13, 2026
- structured_output: true → false
- capabilityvia kilo · Sep 13, 2026
- structured_output: true → false
- capabilityvia amazon-bedrock · Sep 13, 2026
- description: DeepSeek reasoning model for multi-step analysis, math, coding, and tools → Classic open reasoning model for transparent math, coding, and deliberate problem solving
- tool_call: true → false
- open_weights: false → true
- capabilityvia edenai · Sep 7, 2026
- name: DeepSeek-R1 → DeepSeek-R1 (Deep Infra)
Benchmarks
- Aider Polyglot ↗(percent correct)report ↗Independent56.9
- Artificial Analysis Coding Index ↗(index)report ↗Marketplace15.9
- SciCode(percent correct)report ↗Marketplace35.7
- Terminal-Bench Hard ↗(success rate)report ↗Marketplace6.1
Scores a lab published about its own model are claims, not measurements. Rankings elsewhere on this site use independently measured scores only.
Weights
Standard open-source terms. Commercial use without additional conditions. Weights download without an access request.
licence and gating from the Hugging Face model card ↗ — the card is authoritative, this is a summary
Prices are per million tokens (USD). “—” means the provider does not publicly list a price for this model. Data from models.dev; verify with the provider before purchasing.