NanoGPT
currentpqhaz/DeepSeek-V4.1-Flash-1017-heretic:thinking
Input
$0.18Output
$0.72Cache
$0.012Context
512KMax output
128K- Reasoning
- Yes
- Tool calling
- Yes
- Structured output
- No
- Temperature
- Unknown
Recent price
← ModelsCompare model Estimate your workload cost
DeepSeek · current
V4.1 Flash Heretic Thinking
V4.1 Flash Heretic Thinking is attributed to DeepSeek in the catalog. It accepts text and image input and produces text output. Reported capabilities include reasoning and tool calling.
Modalities
Hosted provider token rates
$0.18 / $0.72USD per 1M input / output tokens · NanoGPT
Context
512KReleased
2026-10-31Knowledge cutoff
UnknownNo recent lab-operated provider rate is available. Showing the third-party host with the lowest input rate + ¼ output rate. This is one offering, not an average; other providers, modalities, tiers, and fees may differ.
Where to use it
1 observed offerings · token prices in USD per million where available
Limits and capabilities reflect each provider’s listed offering. Unknown means support was not reported.
Capabilities and limits
- Input
- text, image
- Output
- text
- Max output
- 128K
- Weights
- Open (source-reported)
- Release
- 2026-10-31
- Knowledge cutoff
- Unknown
- Family
- deepseek
Reasoning · Tool calling · Vision
Provider-specific limits and support can differ.
Plan a model evaluation
Evaluate quality and latency on examples from your intended use before choosing a model.
Explore available evidenceModels to compare
Candidates match reported specifications; they are not tested replacements.
Model and price history
Price history (1 observations)
| Date | Provider | Input / output |
|---|---|---|
| 2026-11-01 | NanoGPT | $0.18 / $0.72 |