Gemini 3.5 Flash Lite vs Kimi K2 Instruct vs Qwen3.6 35B-A3B
Gemini 3.5 Flash Lite comes out ahead, 71 to 65 and 53 on our weighted score.
Price in rating: Included · Exclude price
- Our pick
Google
Gemini 3.5 Flash Lite
71/100- ECI145.7
- Price$0.30 / $2.50
- Context1.05M
Moonshot AI
Kimi K2 Instruct
53/100- ECI141.3
- Price$0.57 / $2.30
- Context131K
Alibaba (Qwen)
Qwen3.6 35B-A3B
65/100- ECI144.1
- Price$0.375 / $2.25
- Context262K
Gemini 3.5 Flash Lite is our pick
Gemini 3.5 Flash Lite is the better all-round choice, scoring 71/100 against Qwen3.6 35B-A3B (65) and Kimi K2 Instruct (53). It leads on inputs & features and context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityGemini 3.5 Flash LiteCapabilities Index (ECI): Gemini 3.5 Flash Lite 145.7 · Qwen3.6 35B-A3B 144.1 · Kimi K2 Instruct 141.3
- Lowest priceQwen3.6 35B-A3BQwen3.6 35B-A3B $0.844 · Gemini 3.5 Flash Lite $0.85 · Kimi K2 Instruct $1.00 per 1M tokens (3:1 blend)
- Longest contextGemini 3.5 Flash LiteGemini 3.5 Flash Lite 1,048,576 · Qwen3.6 35B-A3B 262,144 · Kimi K2 Instruct 131,072 tokens
- Widest inputsGemini 3.5 Flash LiteGemini 3.5 Flash Lite: Text, Images, PDFs, Audio, Video · Kimi K2 Instruct: Text · Qwen3.6 35B-A3B: Text, Images, Video
- Self-hostingKimi K2 Instruct and Qwen3.6 35B-A3BPublishes downloadable weights (Modified MIT)
| Measure | Weight | Gemini 3.5 Flash Lite | Kimi K2 Instruct | Qwen3.6 35B-A3B |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 73 | 69 | 72 |
| Price | 25% | 51 | 48 | 51 |
| Inputs & features | 15% | 100 | 25 | 80 |
| Context window | 10% | 65 | 32 | 43 |
| Overall | 100% | 71/100 | 53/100 | 65/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 145.7 (best) | 141.3 | 144.1 |
| ECI rank | #89 of 158 (best) | #108 of 158 | #95 of 158 |
| GPQA DiamondGraduate-level science questions | 83.3% | — | 84.9% (best) |
| FrontierMath Tiers 1–3Research-level mathematics | 26.0% (best) | — | 20.4% |
| OTIS Mock AIME 2024–2025Competition mathematics | 71.1% | — | 86.7% (best) |
| Price per million tokens | |||
| Input | $0.30 (best) | $0.57 | $0.375 |
| Output | $2.50 | $2.30 | $2.25 (best) |
| Cached input | $0.03 | — | — |
| Blended (3:1) | $0.85 | $1.00 | $0.844 (best) |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Official Google API | Median of 11 providers | Official Alibaba API |
| Limits | |||
| Context window | 1,048,576 tokens (best) | 131,072 tokens | 262,144 tokens |
| Max output | 65,536 tokens (best) | 32,768 tokens | 65,536 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | No | Yes |
| PDFs | Yes | No | No |
| Audio | Yes | No | No |
| Video | Yes | No | Yes |
| Reasoning | Yesminimal · low · medium · high | No | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | No | Yes |
| Availability | |||
| Weights | Proprietary | OpenModified MIT | Open |
| API model ID | gemini-3.5-flash-lite | — | qwen3.6-35b-a3b |
| API providers | 33 | 11 | 43 (best) |
| Released | Jul 21, 2026 | Jul 11, 2025 | Apr 17, 2026 |
| Knowledge cutoff | Mar 2026 | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
Gemini 3.5 Flash Lite$8.00
Kimi K2 Instruct$10.30
Qwen3.6 35B-A3B$8.25
Which should you choose?
Which is better: Gemini 3.5 Flash Lite, Kimi K2 Instruct or Qwen3.6 35B-A3B?
Gemini 3.5 Flash Lite is the better all-round choice, scoring 71/100 against Qwen3.6 35B-A3B (65) and Kimi K2 Instruct (53). It leads on inputs & features and context window. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, Gemini 3.5 Flash Lite, Kimi K2 Instruct or Qwen3.6 35B-A3B?
Qwen3.6 35B-A3B is cheaper at $0.375 input / $2.25 output per million tokens (official Alibaba API price). Gemini 3.5 Flash Lite costs $0.30 input / $2.50 output per million tokens (official Google API price); Kimi K2 Instruct costs $0.57 input / $2.30 output per million tokens (median across 11 API providers). At a typical mix of three input tokens to one output token, that is $0.844 per million tokens for Qwen3.6 35B-A3B versus $0.85 for Gemini 3.5 Flash Lite (1× as much) and $1.00 for Kimi K2 Instruct (1.2× as much).
Which scores higher on benchmarks?
Gemini 3.5 Flash Lite scores higher on the Capabilities Index (ECI): Gemini 3.5 Flash Lite 145.7 (#89 of 158), Qwen3.6 35B-A3B 144.1 (#95 of 158) and Kimi K2 Instruct 141.3 (#108 of 158). The confidence ranges of the top two overlap (144.0–147.3 vs 142.2–146.0), so treat the gap as small.
Which is better for coding?
There are no published SWE-bench Verified results for Gemini 3.5 Flash Lite, Kimi K2 Instruct and Qwen3.6 35B-A3B yet, so there is no like-for-like coding score. On overall capability, Gemini 3.5 Flash Lite leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.
Which has the bigger context window?
Gemini 3.5 Flash Lite has the largest context window at 1,048,576 tokens, against 262,144 for Qwen3.6 35B-A3B and 131,072 for Kimi K2 Instruct. Maximum output per response: Gemini 3.5 Flash Lite up to 65,536, Kimi K2 Instruct up to 32,768, Qwen3.6 35B-A3B up to 65,536 tokens.
Which can read images, PDFs, audio or video?
Gemini 3.5 Flash Lite accepts text, images, PDFs, audio and video; Kimi K2 Instruct accepts text; Qwen3.6 35B-A3B accepts text, images and video. Gemini 3.5 Flash Lite handles the widest range of inputs.
Are any of these open source?
Kimi K2 Instruct and Qwen3.6 35B-A3B publishes its weights (Modified MIT) and can be self-hosted; Gemini 3.5 Flash Lite is proprietary.
Which is newer?
Gemini 3.5 Flash Lite is the newest, released Jul 21, 2026. Qwen3.6 35B-A3B came out Apr 17, 2026; Kimi K2 Instruct came out Jul 11, 2025. Knowledge cutoff: Gemini 3.5 Flash Lite Mar 2026.
How do you decide the winner?
This rating uses capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 18, 2026.