Provider comparison · Updated 12 Oct 2026
Anthropic vs DeepSeek
18 models vs 13 models
DeepSeek is cheaper at the median: $0.27 against $5.00 per 1M input tokens across paid models, about 19× less. Its cheapest paid model, DeepSeek V4 Flash 0731 ($0.08), compares with Claude Haiku 5.5 ($0.10) at Anthropic. Anthropic lists 18 models; DeepSeek lists 13 models.
Ask an AI about this pageGChatGPTCClaudePPerplexityGeGeminiXGrok
Cite
Cite this page
ModelsAtlas. “Anthropic vs DeepSeek”. Retrieved 12 Oct 2026. https://modelsatlas.com/compare-providers/anthropic/deepseek
At a glance
Standard tier, paid models · highlighted = better
Anthropic
DeepSeek
Models listed
18
13
Cheapest input / 1M
$0.10
$0.08
Median in / out
$5.00 / $25.00
$0.27 / $0.95
Highest input / 1M
$15.00
$0.70
Largest context
1M
1.05M
Batch-tier models
15 of 18
1 of 13
Open-weight models
0
3
Free models
0
0
Vision models listed
0
1
Costs for real workloads
Cheapest paid model from each provider that fits the job · monthly
Derived from standard-tier prices, 30 days. “Fits the job” means the context window holds the prompt and reply. Excludes batch and caching.
Cheapest model for each capability
Input / output per 1M
| Capability | Anthropic | DeepSeek |
|---|---|---|
| 1M context | Claude Haiku 5.5 $0.10 / $0.50 | DeepSeek V4 Flash 0731 $0.08 / $0.17 |
Line-up by price tier
Paid models · input per 1M
| Tier | Anthropic | DeepSeek |
|---|---|---|
| Flagship $3+ | Claude Opus 4.1 $15.00 Claude Opus 4 $15.00 Claude Fable 5.1 $10.00 Claude Fable 5 $10.00 Claude Opus 5 $5.00 +8 more | None listed |
| Mid $0.50–$3 | Claude Haiku 4.5 $1.00 Claude Sonnet 5 $2.00 Claude Sonnet 5.5 $2.00 | R1 0528 $0.70 |
| Budget under $0.50 | Claude Haiku 5.5 $0.10 | DeepSeek V4 Flash 0731 $0.08 DeepSeek V4 Flash Vision Exp $0.22 DeepSeek V3.1 $0.25 DeepSeek V3.2 $0.26 DeepSeek V3.1 Terminus $0.27 +3 more |
| Open weights any price | None listed | None listed |
Questions
Is Anthropic or DeepSeek cheaper?
DeepSeek at the median: $0.27 against $5.00 for Anthropic per 1M input tokens. Median is across each provider's paid standard-tier models, so a provider with many small models can look cheaper than its flagship is.
Which provider has more models?
Anthropic lists more: Anthropic 18, DeepSeek 13. Counts are models with a page in our catalogue.
Do both offer batch pricing?
Anthropic: 15 models list a batch tier. DeepSeek: 1 model list a batch tier. Where a model lists both, batch is a median 50% of standard for Anthropic and 37% for DeepSeek.
Which has the larger context window?
DeepSeek has the larger: Claude Fable 5 (1M tokens) at Anthropic, DeepSeek V4 Flash 0731 (1.05M) at DeepSeek.
Does either provider offer open-weight models?
Anthropic: none marked open-weight. DeepSeek: 3 open-weight models.
Does either have free models?
Anthropic: none listed at no per-token charge. DeepSeek: none.
Browse more providers
Keep going in your AI
Opens with this page as the source. We add nothing else to the prompt.
- Is Anthropic or DeepSeek cheaper?
- Which provider has more models?
- Do both offer batch pricing?