Multimodal Kimi model with 1M context and toggleable max-effort thinking for long-horizon agent work
Benchmarks
| Benchmark | Category | Score | Variant | Provenance | Source | Observed |
|---|---|---|---|---|---|---|
| AA Agentic Index | agentic | 50.0 | — | mirror | openrouter | Sep 20, 2026 |
| AA Coding Index | coding | 76.2 | — | mirror | openrouter | Jul 23, 2026 |
| AA Intelligence Index | intelligence | 43.6 | — | mirror | openrouter | Sep 20, 2026 |
| Design Arena — 3d | preference | 1425 | — | mirror | openrouter | Sep 15, 2026 |
| Design Arena — agenticgamedev | preference | 1252 | — | mirror | openrouter | Sep 12, 2026 |
| Design Arena — androidnative | preference | 1259 | — | mirror | openrouter | Sep 21, 2026 |
| Design Arena — codecategories | preference | 1384 | — | mirror | openrouter | Sep 21, 2026 |
| Design Arena — dataviz | preference | 1358 | — | mirror | openrouter | Sep 20, 2026 |
| Design Arena — fullstack | preference | 1330 | — | mirror | openrouter | Sep 21, 2026 |
| Design Arena — gamedev | preference | 1407 | — | mirror | openrouter | Sep 20, 2026 |
| Design Arena — godotgamedev | preference | 1199 | — | mirror | openrouter | Jul 23, 2026 |
| Design Arena — htmlslides | preference | 1259 | — | mirror | openrouter | Sep 9, 2026 |
| Design Arena — mobileapps | preference | 1280 | — | mirror | openrouter | Sep 20, 2026 |
| Design Arena — python-pptxslides | preference | 1273 | — | mirror | openrouter | Sep 11, 2026 |
| Design Arena — svg | preference | 1330 | — | mirror | openrouter | Sep 19, 2026 |
| Design Arena — uicomponent | preference | 1374 | — | mirror | openrouter | Sep 21, 2026 |
| Design Arena — webapps | preference | 1309 | — | mirror | openrouter | Sep 16, 2026 |
| Design Arena — website | preference | 1352 | — | mirror | openrouter | Sep 20, 2026 |
| LMArena Text | preference | 1482 | — | crowd | lmarena | Jul 21, 2026 |
| LMArena WebDev | preference | 1682 | — | crowd | lmarena | Jul 24, 2026 |
Includes index data from Artificial Analysis.
Providers
Live operational stats per hosting endpoint, via OpenRouter.
| Provider | Quant | Context | $/1M in · out | Uptime 24h | Latency | Throughput |
|---|---|---|---|---|---|---|
| Alibaba | — | 1.0M | $3.45 · $17.25 | 99.21% | — | — |
| BaseTen | fp8 | 1.0M | $3 · $15 | 99.82% | — | — |
| Chutes | mxfp4 | 1.0M | $3 · $15 | 97.21% | — | — |
| DeepInfra | bf16 | 1.0M | $2.85 · $14.25 | 95.70% | — | — |
| DigitalOcean | — | 1.0M | $2.55 · $12.95 | 99.70% | — | — |
| Fireworks | — | 1.0M | $3 · $15 | 99.74% | — | — |
| Fireworks | — | 1.0M | $4.5 · $22.5 | 99.12% | — | — |
| Fireworks | — | 1.0M | $3.3 · $16.5 | 99.40% | — | — |
| InferenceNet | fp4 | 1.0M | $2.1 · $10.95 | 99.21% | — | — |
| Makora | — | 1.0M | $2.55 · $12.75 | 96.57% | — | — |
| Modal | mxfp4 | 1.0M | $3 · $15 | 99.49% | — | — |
| Moonshot AI | mxfp4 | 1.0M | $3 · $15 | 99.95% | — | — |
| Morph | — | 1.0M | $2.13 · $11.9 | 99.61% | — | — |
| Morph | — | 1.0M | $6 · $22.5 | 99.84% | — | — |
| Parasail | fp4 | 1.0M | $3 · $15 | 98.33% | — | — |
| Phala | — | 1.0M | $2.4 · $12 | 95.03% | — | — |
| Relace | fp4 | 1.0M | $1.7 · $8.5 | 99.72% | — | — |
| Sail Research | fp4 | 1.0M | $2.14 · $10.76 | 99.69% | — | — |
| Together | — | 1.0M | $3 · $15 | 99.82% | — | — |
| Wafer | — | 1.0M | $2.1 · $10.95 | 99.94% | — | — |