DeepSeek V4.1 Flash vs Ternary Bonsai 4B
At a Glance
| Compare |  DeepSeek V4.1 FlashDeepSeek |  Ternary Bonsai 4BPrismML |
|---|---|---|
| Intelligence, Cost, and Efficiency | ||
| CostLower is better · Published-token output estimate | #5 of 44$0.022 per LiveBench case | UnrankedNot in the 44-model eligible cohort |
| Pricing and Limits | ||
| Input priceFrom · USD / 1M tokens | $0.15DeepSeek ↗ · Sep 10, 2026 | Not reported |
| Output priceFrom · USD / 1M tokens | $0.60Deepinfra ↗ · Sep 22, 2026 | Not reported |
| Context windowMaximum documented tokens | 1,049K | 33K |
| Model facts checked | Sep 10, 2026View model evidence → | Sep 18, 2026View model evidence → |
Token prices are the lowest available sourced USD rates; input and output may use different providers. Cost ranking estimates output spend on LiveBench, not a full request bill. Ranking methodology →
Available Benchmarks
Side-by-Side Facts
| Field | DeepSeek-V4.1-Flash | Ternary Bonsai 4B |
|---|---|---|
| Developer | DeepSeek | PrismML |
| Family | Deepseek V4 1 | Bonsai 4b |
| Model | DeepSeek-V4.1-Flash | Ternary Bonsai 4B |
| Version | DeepSeek-V4.1-Flash | Ternary Bonsai 4B |
| Lifecycle | active | active |
| Released | 2026-09-10 | 2026-04-18 |
| Knowledge cutoff | Unknown | Unknown |
| Input modalities | Text, Image | Text |
| Output modalities | Text | Text |
| Context window | 1,049K | 33K |
| Total parameters | 763.2B | 4B |
| Active parameters | Unknown | Unknown |
| License | mit | apache-2.0 |
| Open weights | Yes | Yes |
| API available | Yes | No |
| Self-hostable | Yes | Yes |
| Provider access | DeepSeek (Standard), Deepinfra (Standard), Together Ai (Standard) | Unknown |
| Capabilities | agents, chat, fim, generation, reasoning, responses, structured_outputs, tools, vision | chat, generation |
| Architecture design | Causal Encoder-Decoder (20 encoder + 20 decoder layers) | Unknown |
| Backbone parameters | 552000000000 parameters | Unknown |
| Active parameters during decode | 16000000000 parameters | Unknown |
| Effective bit width | Unknown | 1.58 bits per weight |
| Active parameters during prefill | 8000000000 parameters | Unknown |
| Pre-training corpus | 45000000000000 tokens | Unknown |
| Reasoning effort range | 1–100 | Unknown |
| Routed experts per MoE layer | 384 experts | Unknown |
| Routed experts per token | 6 experts | Unknown |
| Transformer layers | 40 layers | Unknown |
| Weight size | Unknown | 1.07 GB |
| Weight format | Unknown | Ternary Q2_0 |
DeepSeek V4.1 Flash Capabilities
Ternary Bonsai 4B Capabilities
Related Comparisons
All text comparisons →| A | Pair | B | Context |
|---|---|---|---|
DeepSeek-V4-FlashDeepSeek | vs | DeepSeek-V4.1-FlashDeepSeek | family variantstext |
DeepSeek-V4.1-FlashDeepSeek | vs |  GPT-6 AstraOpenAI | cross-developer peersimage, text |
Claude Fable 5.1Anthropic | vs | DeepSeek-V4.1-FlashDeepSeek | cross-developer peersimage, text |
DeepSeek-V4.1-FlashDeepSeek | vs |  Gemini 3.1 ProGoogle DeepMind | cross-developer peerstext |
DeepSeek-V4.1-FlashDeepSeek | vs | Grok 4.6xAI | cross-developer peersimage, text |
DeepSeek-V4.1-FlashDeepSeek | vs |  Qwen3.8-MaxQwen | cross-developer peersimage, text |
DeepSeek-V4.1-FlashDeepSeek | vs |  Kimi-K3Moonshot AI | cross-developer peersimage, text |
 MiniMax-M3MiniMax | vs | DeepSeek-V4.1-FlashDeepSeek | cross-developer peersimage, text |
DeepSeek-V4.1-FlashDeepSeek | vs |  GLM-5.3Z.ai | cross-developer peerstext |
DeepSeek-V4.1-FlashDeepSeek | vs |  Hy4 previewTencent | cross-developer peerstext |
Seed 2.1 ProByteDance Seed | vs | DeepSeek-V4.1-FlashDeepSeek | cross-developer peersimage, text |
DeepSeek-V4.1-FlashDeepSeek | vs |  Mistral Large 3Mistral AI | cross-developer peersimage, text |
Primary Evidence
Sources and Freshness
Questions
DeepSeek V4.1 Flash vs Ternary Bonsai 4B FAQs
Is DeepSeek V4.1 Flash or Ternary Bonsai 4B better for coding?+
This comparison does not currently contain a protocol-matched coding benchmark for both DeepSeek V4.1 Flash and Ternary Bonsai 4B, so Model Markets cannot name a coding leader from pricing, context size, or capability labels alone.
Which is cheaper, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
Only DeepSeek V4.1 Flash has a directly sourced input price: $0.15 per million tokens. Only DeepSeek V4.1 Flash has a directly sourced output price: $0.60 per million tokens.
Which has a larger context window, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
DeepSeek V4.1 Flash has the larger sourced context window. DeepSeek V4.1 Flash supports 1,049K and Ternary Bonsai 4B supports 33K.
Which performs better in benchmarks, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
There is no overall benchmark winner: At least two independently verified, protocol-matched benchmarks are required for an overall winner.
Can DeepSeek V4.1 Flash or Ternary Bonsai 4B be self-hosted?+
Both models have the same recorded self-hosting status: supported. DeepSeek V4.1 Flash is open weight; Ternary Bonsai 4B is open weight.
Can DeepSeek V4.1 Flash and Ternary Bonsai 4B understand images?+
DeepSeek V4.1 Flash is documented with image input; Ternary Bonsai 4B is not documented with image input. This reflects supported input modalities, not vision quality.
Which can generate longer answers, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
Neither has a larger sourced maximum output. DeepSeek V4.1 Flash is 393K and Ternary Bonsai 4B is —.
Do DeepSeek V4.1 Flash and Ternary Bonsai 4B support reasoning and tool use?+
DeepSeek V4.1 Flash: reasoning, tool calling, and image input. Ternary Bonsai 4B: none of these features are definitively sourced. Feature support does not establish relative quality.
Which is available from more inference providers, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
DeepSeek V4.1 Flash has 3 sourced provider routes; Ternary Bonsai 4B has 0, so DeepSeek V4.1 Flash has broader tracked availability.
Which offers better value, DeepSeek V4.1 Flash or Ternary Bonsai 4B?+
There is no universal value winner. Compare the input and output prices above with the matched benchmark result for your workload: cheaper tokens can be offset by different quality, token usage, latency, or provider availability.