Kimi K3.1 Pro vs GPT 5.6 Sol
Kimi K3.1 Pro is from Moonshot AI, released October 24, 2026. GPT 5.6 Sol is from OpenAI, released July 9, 2026.
- LMArena Elo1485 vs 1485. Both rated 1485 on LMArena.
- Input per 1M$2 vs $4. Kimi K3.1 Pro costs 50% less per million input tokens.
- Output per 1M$10 vs $20. Kimi K3.1 Pro costs 50% less per million output tokens.
- Context window1M vs 1.05M. GPT 5.6 Sol takes 5.0% more context in one request.
Prices are the vendors’ published API rates; Elo is LMArena’s text board. Each figure is dated on its model page.
Interactive Comparison
Loading chart...
Token cost calculator
Input tokens
Output tokens
| Model | Input | Output | Cost for this many tokens |
|---|---|---|---|
| Kimi K3.1 Pro | $0.20 | $0.20 | $0.40 |
| GPT 5.6 Sol | $0.40 | $0.40 | $0.80 |
List prices per 1M tokens from each model's record (Kimi K3.1 Pro: $2 in / $10 out; GPT 5.6 Sol: $4 in / $20 out). Cached-input and batch discounts are not included. Click a figure to type it. Compare up to 8 models in the token cost calculator.
$2 / $10 per million tokens | $4 / $20 per million tokens | |
|---|---|---|
| Identity | ||
| Developer | Moonshot AI | OpenAI |
| Released | Oct 2026 | Jul 2026 |
| Status | GA | — |
| Licence | Open weights on Hugging Face (moonshotai/Kimi-K3.1-Pro) under the Kimi K3.1 Pro licence | Proprietary |
| Self-hostable | Not yet — weights due by 4 November 2026 | No |
| Cost | ||
| Blended $/1M tokensinput × 0.75 + output × 0.25 | $4.000 / 1M tokens | $8.000 / 1M tokens |
| Input price | $2.000 / 1M tokens | $4.000 / 1M tokens |
| Output price | $10.000 / 1M tokens | $20.000 / 1M tokens |
| Cached input | $0.20 / 1M tokens; cache writes $2 (5-minute TTL) or $4 (1-hour TTL) | 0.40 |
| Free tier | Not offered — unlocked after a minimum $1 top-up | — |
| Capacity | ||
| Context window | 1M tokens | 1.05M tokens |
| Max output | 131,072 tokens by default; max_completion_tokens up to 1,048,576 | 128,000 tokens |
| Long-context surcharge | — | Prompts over 272K input tokens bill at 2x input and 1.5x output for the whole request |
| Capability | ||
| Vision in | Yes — images and video (base64 or uploaded files; no public URLs) | — |
| Audio in | Not stated | — |
| Function calling | Yes | — |
| Structured output | Yes — json_schema with strict: true | — |
| Extended reasoning | Yes — always on; reasoning_effort low, high or max (default max) | — |
| Web search | Being updated by Moonshot; not recommended for production in the near term | — |
| Access | ||
| API | Yes | — |
| Chat app | Yes | — |
| Measured quality | ||
| LMArena Elo | 1485 (checked Nov 2026) | 1485 (checked Nov 2026) |
| LMArena rank | Rank 24 | Rank 23 |
| Elo per dollarLMArena Elo ÷ blended $/1M tokens | 371 | 186 |
| LMArena by task | ||
| Coding | 1530 Elo (rank 23) | 1533 Elo (rank 19) |
| Math | — | 1494 Elo (rank 23) |
| Creative Writing | 1453 Elo (rank 36) | 1471 Elo (rank 19) |
| Instruction Following | 1487 Elo (rank 18) | 1490 Elo (rank 16) |
| Hard Prompts | 1504 Elo (rank 29) | 1511 Elo (rank 19) |
| Longer Query | 1491 Elo (rank 29) | 1498 Elo (rank 21) |
| Multi-Turn | 1475 Elo (rank 53) | 1491 Elo (rank 23) |
| Expert | 1515 Elo (rank 25) | 1532 Elo (rank 12) |
