Lyria 3.5 vs Kimi K2 Thinking
At a Glance
| Compare |  Lyria 3.5Google DeepMind |  Kimi K2 ThinkingMoonshot AI |
|---|---|---|
| Pricing and Limits | ||
| Input priceFrom · USD / 1M tokens | Not reported | $0.60Openrouter ↗ · Sep 22, 2026 |
| Output priceFrom · USD / 1M tokens | Not reported | $2.50Openrouter ↗ · Sep 22, 2026 |
| Context windowMaximum documented tokens | 1,049K | 262K |
| Model facts checked | Sep 22, 2026View model evidence → | Aug 28, 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 | Lyria 3.5 | Kimi-K2-Thinking |
|---|---|---|
| Developer | Google DeepMind | Moonshot AI |
| Family | Lyria | Kimi K2 Thinking |
| Model | Lyria 3.5 | Kimi-K2-Thinking |
| Version | Lyria 3.5 | Kimi-K2-Thinking |
| Lifecycle | active | active |
| Released | 2026-07-29 | 2025-11-06 |
| Knowledge cutoff | Unknown | Unknown |
| Input modalities | Text, Image | Text |
| Output modalities | Text, Audio | Text |
| Context window | 1,049K | 262K |
| Total parameters | Unknown | 1T |
| Active parameters | Unknown | 32B |
| License | Unknown | other |
| Open weights | No | Yes |
| API available | Yes | Yes |
| Self-hostable | No | Yes |
| Provider access | Google AI (Standard), Google Gemini (Standard) | Hugging Face (Standard), Openrouter (Standard) |
| Capabilities | generation, image-to-music, lyrics, music-generation, synthid, vocals | chat, generation, reasoning, tools |
Lyria 3.5 Capabilities
Kimi K2 Thinking Capabilities
Related Comparisons
All text comparisons →| A | Pair | B | Context |
|---|---|---|---|
Lyria 3.5Google DeepMind | vs |  GPT-6 AstraOpenAI | cross-developer peersimage, text |
Claude Fable 5.1Anthropic | vs | Lyria 3.5Google DeepMind | cross-developer peersimage, text |
Gemini 3.1 ProGoogle DeepMind | vs | Lyria 3.5Google DeepMind | developer peerstext |
 DeepSeek-V4.1-FlashDeepSeek | vs | Lyria 3.5Google DeepMind | cross-developer peersimage, text |
Lyria 3.5Google DeepMind | vs | Grok 4.6xAI | cross-developer peersimage, text |
Lyria 3.5Google DeepMind | vs |  Qwen3.8-MaxQwen | cross-developer peersimage, text |
Lyria 3.5Google DeepMind | vs | Kimi-K3Moonshot AI | cross-developer peersimage, text |
 MiniMax-M3MiniMax | vs | Lyria 3.5Google DeepMind | cross-developer peersimage, text |
Lyria 3.5Google DeepMind | vs |  GLM-5.3Z.ai | cross-developer peerstext |
Lyria 3.5Google DeepMind | vs |  Hy4 previewTencent | cross-developer peerstext |
Seed 2.1 ProByteDance Seed | vs | Lyria 3.5Google DeepMind | cross-developer peersimage, text |
Lyria 3.5Google DeepMind | vs |  Mistral Large 3Mistral AI | cross-developer peersimage, text |
Primary Evidence
Sources and Freshness
Questions
Lyria 3.5 vs Kimi K2 Thinking FAQs
Is Lyria 3.5 or Kimi K2 Thinking better for coding?+
This comparison does not currently contain a protocol-matched coding benchmark for both Lyria 3.5 and Kimi K2 Thinking, so Model Markets cannot name a coding leader from pricing, context size, or capability labels alone.
Which is cheaper, Lyria 3.5 or Kimi K2 Thinking?+
Only Kimi K2 Thinking has a directly sourced input price: $0.60 per million tokens. Only Kimi K2 Thinking has a directly sourced output price: $2.50 per million tokens.
Which has a larger context window, Lyria 3.5 or Kimi K2 Thinking?+
Lyria 3.5 has the larger sourced context window. Lyria 3.5 supports 1,049K and Kimi K2 Thinking supports 262K.
Which performs better in benchmarks, Lyria 3.5 or Kimi K2 Thinking?+
There is no overall benchmark winner: At least two independently verified, protocol-matched benchmarks are required for an overall winner.
Can Lyria 3.5 or Kimi K2 Thinking be self-hosted?+
Kimi K2 Thinking is the only model in this pair currently marked as self-hostable. Lyria 3.5 is not marked open weight; Kimi K2 Thinking is open weight.
Can Lyria 3.5 and Kimi K2 Thinking understand images?+
Lyria 3.5 is documented with image input; Kimi K2 Thinking is not documented with image input. This reflects supported input modalities, not vision quality.
Which can generate longer answers, Lyria 3.5 or Kimi K2 Thinking?+
Kimi K2 Thinking has the larger sourced maximum output: Lyria 3.5 supports 66K and Kimi K2 Thinking supports 131K output tokens.
Do Lyria 3.5 and Kimi K2 Thinking support reasoning and tool use?+
Lyria 3.5: image input. Kimi K2 Thinking: reasoning and tool calling. Feature support does not establish relative quality.
Which is available from more inference providers, Lyria 3.5 or Kimi K2 Thinking?+
Lyria 3.5 has 2 sourced provider routes; Kimi K2 Thinking has 2, a tie.
Which offers better value, Lyria 3.5 or Kimi K2 Thinking?+
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.