Kimi K2.6 vs GPT-5.5 vs Gemini 3.1 Pro
Pick two to four models to compare benchmark scores, prices, context windows and speed. The link updates as you pick, so you can share the exact comparison.
Cheapest blended
Kimi K2.6
$1.71 / 1M tokens
Largest context
GPT-5.5
1.05M tokens
Fastest output
Gemini 3.1 Pro
114 tokens/s
Top on SWE-bench Verified
Gemini 3.1 Pro
80.6%
Specifications
| Spec | Kimi K2.6Moonshot | GPT-5.5OpenAI | Gemini 3.1 ProGoogle |
|---|---|---|---|
| Provider | Moonshot | OpenAI | |
| Released | Apr 2026 | 24 Apr 2026 | 19 Feb 2026 |
| Licence | Open weight | Proprietary | ProprietaryPreview |
| Context window | 262K tokens | 1.05M tokens | 1.05M tokens |
| Max output | – | 128K tokens | 66K tokens |
| Input / 1M | $0.95* | $5.00* | $2.00* |
| Output / 1M | $4.00 | $30.00 | $12.00 |
| Cached input / 1M | $0.16 | $0.50 | $0.20 |
| Speed | 68 tokens/s | 86 tokens/s | 114 tokens/s |
| Input types | text, image, video | text, image | text, image, video, audio, pdf |
| Best for | CodingAgentsMultimodal | CodingAgents | ReasoningCodingAgentsMultimodal |
* Kimi K2.6: Prices exclude taxes. Cache hits are automatic.
* GPT-5.5: There is no separate cache-write charge (pre-GPT-5.6 models). Batch and Flex: $2.50 / $15. Fast mode: 2.5x, $12.50 / $75. Snapshot gpt-5.5-2026-04-23. GPT-5.5 Pro (gpt-5.5-pro) costs $30 / $180. Data residency +10%.
* Gemini 3.1 Pro: Still preview. A separate endpoint, gemini-3.1-pro-preview-customtools, has the same price. Cache storage $4.50 per 1M tokens per hour. Priority is $3.60 / $21.60 up to 200K.
Benchmark scores
Officially reported scores only. A model with no published score on a benchmark shows a line, not a zero.
Price per 1M tokens
Context window and speed
Notes
- Kimi K2.6
- Moonshot's open multimodal agent model, 1T total and 32B active parameters, for long-horizon coding and agent swarms. Modified MIT license: commercial use allowed, but products over 100M MAU or $20M monthly revenue must display "Kimi K2.6".
- GPT-5.5
- A frontier model for coding and complex professional work, the flagship before GPT-5.6.
- Gemini 3.1 Pro
- Google's current Pro model, still preview. Built for multimodal understanding, agentic work and vibe-coding, with a focus on software engineering and multi-step tool use.
Still torn between two?
Test them on your own prompts.
Benchmarks are someone else’s work. On a free call we look at yours and agree how to check quality, cost and latency for the models you are weighing.