Hybrid-reasoning DeepSeek model with thinking and non-thinking modes, sparse attention, and tool-use
Specification
deepseek listing
Context
66K
Max output
8K
Released
2025-12-01
Knowledge cutoff
2024-07
Retires
2026-07-13
Open weights
yes
Input
text
Output
text
Price
US dollars per million tokens · deepseek list price
Input
$0.27
Output
$1.10
Cache read
$0.07
Cache write
free
Quality
7 benchmarks · 3 sources
Composite
68th
percentile of 334 scored models
Rank
108 / 334
±0.06 sd
Evidence
7 × 3
3 or more benchmarks, from 2 or more sources
Effort range
10.0
points between effort settings
coding71st2/3 bench
reasoning60th2/4 bench
math57th3/7 bench
Every score
one row per source and configuration — nothing averaged away
| Benchmark | Score | Configuration | Source | Run |
|---|---|---|---|---|
| Aider polyglot | 72.2 | default · aider diff · 2 runs | Aider | 2025-10-03 |
| GPQA diamond | 77.3 ±3.2 | default · 2 runs | Epoch AI | 2026-07-16 |
| OTIS Mock AIME 2024-2025 | 68.4 ±7.5 | default · 2 runs | Epoch AI | 2026-07-16 |
| SWE-Bench verified | 70.0 | high · mini-SWE-agent | SWE-bench | — |
| ↳ | 60.0 | default · mini-SWE-agent | SWE-bench | — |
| Settings differ by 10.0 points on this benchmark. | ||||
| FrontierMath-v1 | 22.1 ±4.1 | default | Epoch AI | 2025-12-22 |
| FrontierMath-Tier-4-2025-07-01 | 2.1 ±2.1 | default | Epoch AI | 2025-12-16 |
| Chess Puzzles | 7.5 ±1.0 | default · 2 runs | Epoch AI | 2026-07-16 |
The leading figure for a benchmark is the median across its configurations, so one heroic high-effort run cannot set the number.
Available from 27 hosts
| Host | In $/M | Out $/M | Cache rd | Cache wr | Context | Output | Retires |
|---|---|---|---|---|---|---|---|
| deepseekvendor | $0.27 $0.28 | $1.10 $0.40 | $0.07 | free | 66K 164K | 8K 164K | — |
| hyperbolic | $0.20 $0.40 | $0.20 $0.40 | — | — | 33K | 33K | — |
| lambda_ai | $0.20 | $0.60 | — | — | 131K | 131K | — |
| tokengo | $0.2174 $0.19 | $0.326 $0.71 | $0.06 | — | 128K 131K | 64K 8K | — |
| kilo | $0.269 $0.27 | $0.40 $0.41 | $0.1345 | — | 164K | 66K | — |
| llmgateway-providers | $0.269 $0.26 | $0.40 $0.38 | $0.1345 $0.13 | — | 164K 160K | 66K | — |
| novita | $0.269 $0.27 $0.89 | $0.40 $0.41 $1.00 $1.12 $0.89 | $0.1345 $0.135 | — | 164K 131K 64K | 66K 33K 16K | — |
| novita-ai | $0.269 $0.27 | $0.40 $0.41 $1.12 $1.00 | $0.1345 $0.135 | — | 164K 131K | 66K 164K 33K | — |
| openrouter | $0.269 $0.27 | $0.40 $0.41 | $0.1345 | — | 164K | 66K 164K | — |
| poe | $0.27 | $0.40 | $0.13 | — | 128K | 0 | — |
| jiekou | $0.28 $0.27 | $1.14 $1.00 | — | — | 164K | 164K 33K | — |
| nano-gpt | $0.28 $1.10 | $0.42 $2.50 | $0.14 $0.04 | — | 163K 1M | 66K 384K | — |
| nano-gpt · thinking | $0.28 | $0.42 | $0.14 | — | 163K | 66K | — |
| zenmux | $0.28 $0.22 | $0.43 $0.33 | — | — | 128K 163K | 64K | — |
| ofox | $0.29 | $0.43 | $0.06 | — | 128K | 32K | — |
| tensorx | $0.30 | $0.50 | $0.075 | $0.375 | 164K | 164K | — |
| deepinfra | $0.32 $0.24 $0.25 $0.26 | $0.89 $0.90 $0.95 $0.38 | $0.135 $0.216 $0.13 | — | 164K | 164K | — |
| edenai | $0.32 $0.24 | $0.89 $0.90 | $0.135 | — | 164K | 8K 164K | — |
| baseten | $0.50 $0.77 | $1.50 $0.77 | — | — | — | — | — |
| nebius | $0.50 | $1.50 | — | — | 128K | 128K | — |
| abacus | $0.55 | $1.66 | — | — | 128K | 8K | — |
| azure_ai | $0.58 $1.14 $1.23 | $1.68 $4.56 $4.94 | — | — | 164K 128K 131K | 164K 8K 131K | 2026-07-13 |
| merge-gateway | $0.58 $0.28 $0.50 | $1.68 $0.45 $1.50 | $0.14 | — | 164K 164K | 82K 41K 41K | — |
| vercel | $0.62 $0.25 | $1.85 $0.95 | $0.13 | — | 128K 164K | 8K 128K | — |
| fireworks_ai | $0.90 $0.56 $1.20 | $0.90 $1.68 $1.20 | — | — | 128K 164K 33K | 8K 164K 33K | — |
| together_ai | $1.25 $0.60 | $1.25 $1.70 | — | — | 66K 128K | 8K 16K | 2026-05-14 |
| qiniu-ai | — | — | — | — | 128K | 32K | — |
Price history
append-only observations · a listing writes a row only when its price moves
| Date | Host | Field | $/M |
|---|---|---|---|
| 2026-09-02 | llmgateway-providers | Cache read | $0.13 |
| 2026-09-04 | llmgateway-providers | Cache read | $0.1345 |
| 2026-09-02 | novita-ai | Cache read | $0.135 |
| 2026-09-04 | novita-ai | Cache read | $0.1345 |
| 2026-09-02 | vercel | Cache read | $0.13 |
| 2026-09-04 | vercel | Cache read | $0.028 |
| 2026-09-11 | vercel | Cache read | $0.13 |
| 2026-09-02 | edenai | Input | $0.24 |
| 2026-09-04 | edenai | Input | $0.32 |
| 2026-09-02 | llmgateway-providers | Input | $0.26 |
| 2026-09-04 | llmgateway-providers | Input | $0.269 |
| 2026-09-02 | merge-gateway | Input | $0.50 |
| 2026-09-04 | merge-gateway | Input | $0.58 |
| 2026-09-02 | novita-ai | Input | $0.27 |
| 2026-09-04 | novita-ai | Input | $0.269 |
| 2026-09-02 | openrouter | Input | $0.269 |
| 2026-09-06 | openrouter | Input | $0.269 |
| 2026-09-02 | tokengo | Input | $0.19 |
| 2026-09-04 | tokengo | Input | $0.2174 |
| 2026-09-02 | vercel | Input | $0.25 |
| 2026-09-04 | vercel | Input | $0.28 |
| 2026-09-11 | vercel | Input | $0.62 |
| 2026-09-02 | edenai | Output | $0.90 |
| 2026-09-04 | edenai | Output | $0.89 |
| 2026-09-02 | llmgateway-providers | Output | $0.38 |
| 2026-09-04 | llmgateway-providers | Output | $0.40 |
| 2026-09-02 | merge-gateway | Output | $1.50 |
| 2026-09-04 | merge-gateway | Output | $1.68 |
| 2026-09-02 | novita-ai | Output | $1.12 |
| 2026-09-04 | novita-ai | Output | $0.40 |
| 2026-09-02 | openrouter | Output | $0.40 |
| 2026-09-06 | openrouter | Output | $0.40 |
| 2026-09-02 | tokengo | Output | $0.71 |
| 2026-09-04 | tokengo | Output | $0.326 |
| 2026-09-02 | vercel | Output | $0.95 |
| 2026-09-04 | vercel | Output | $0.42 |
| 2026-09-11 | vercel | Output | $1.85 |
Other listings of this model
same model, different key
More from DeepSeek
most-hosted first