Kimi K2 Instruct vs Qwen3 235B-A22B Thinking 2507 vs Qwen3.5 35B-A3B
Qwen3.5 35B-A3B comes out ahead, 70 to 60 and 55 on our weighted score, and it is the cheaper option too.
Price in rating: Excluded · Include price
Moonshot AI
Kimi K2 Instruct
55/100- ECI141.3
- Price$0.57 / $2.30
- Context131K
Alibaba (Qwen)
Qwen3 235B-A22B Thinking 2507
60/100- ECI143.3
- Price$0.30 / $2.90
- Context262K
- Our pick
Alibaba (Qwen)
Qwen3.5 35B-A3B
70/100- ECI144.1
- Price$0.25 / $2.00
- Context262K
Qwen3.5 35B-A3B is our pick
Qwen3.5 35B-A3B is the better all-round choice, scoring 70/100 against Qwen3 235B-A22B Thinking 2507 (60) and Kimi K2 Instruct (55). It leads on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%.
- CapabilityQwen3.5 35B-A3BCapabilities Index (ECI): Qwen3.5 35B-A3B 144.1 · Qwen3 235B-A22B Thinking 2507 143.3 · Kimi K2 Instruct 141.3
- Lowest priceQwen3.5 35B-A3BQwen3.5 35B-A3B $0.688 · Qwen3 235B-A22B Thinking 2507 $0.95 · Kimi K2 Instruct $1.00 per 1M tokens (3:1 blend)
- Longest contextQwen3 235B-A22B Thinking 2507 and Qwen3.5 35B-A3BQwen3 235B-A22B Thinking 2507 262,144 · Qwen3.5 35B-A3B 262,144 · Kimi K2 Instruct 131,072 tokens
- Widest inputsQwen3.5 35B-A3BKimi K2 Instruct: Text · Qwen3 235B-A22B Thinking 2507: Text · Qwen3.5 35B-A3B: Text, Images, Video
- Self-hostingAll open weightsEvery model here can be downloaded and run on your own hardware
| Measure | Weight | Kimi K2 Instruct | Qwen3 235B-A22B Thinking 2507 | Qwen3.5 35B-A3B |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 69 | 71 | 72 |
| Inputs & features | 20% | 25 | 35 | 80 |
| Context window | 13% | 32 | 43 | 43 |
| Overall | 100% | 55/100 | 60/100 | 70/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 141.3 | 143.3 | 144.1 (best) |
| ECI rank | #108 of 158 | #100 of 158 | #96 of 158 (best) |
| GPQA DiamondGraduate-level science questions | — | 80.1% | 83.5% (best) |
| OTIS Mock AIME 2024–2025Competition mathematics | — | 86.7% (best) | 70.0% |
| SimpleQA VerifiedShort factual questions | — | 40.4% | — |
| Price per million tokens | |||
| Input | $0.57 | $0.30 | $0.25 (best) |
| Output | $2.30 | $2.90 | $2.00 (best) |
| Cached input | — | — | — |
| Blended (3:1) | $1.00 | $0.95 | $0.688 (best) |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 11 providers | Median of 11 providers | Official Alibaba API |
| Limits | |||
| Context window | 131,072 tokens | 262,144 tokens (best) | 262,144 tokens (best) |
| Max output | 32,768 tokens | 81,920 tokens (best) | 65,536 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | No | Yes |
| PDFs | No | No | No |
| Audio | No | No | No |
| Video | No | No | Yes |
| Reasoning | No | Yes | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | No | Yes |
| Availability | |||
| Weights | OpenModified MIT | OpenApache 2.0 | Open |
| API model ID | — | — | qwen3.5-35b-a3b |
| API providers | 11 | 12 | 24 (best) |
| Released | Jul 11, 2025 | Jul 25, 2025 | Feb 23, 2026 |
| Knowledge cutoff | — | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
Kimi K2 Instruct$10.30
Qwen3 235B-A22B Thinking 2507$8.80
Qwen3.5 35B-A3B$6.50
Which should you choose?
Which is better: Kimi K2 Instruct, Qwen3 235B-A22B Thinking 2507 or Qwen3.5 35B-A3B?
Qwen3.5 35B-A3B is the better all-round choice, scoring 70/100 against Qwen3 235B-A22B Thinking 2507 (60) and Kimi K2 Instruct (55). It leads on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, Kimi K2 Instruct, Qwen3 235B-A22B Thinking 2507 or Qwen3.5 35B-A3B?
Qwen3.5 35B-A3B is cheaper at $0.25 input / $2.00 output per million tokens (official Alibaba API price). Qwen3 235B-A22B Thinking 2507 costs $0.30 input / $2.90 output per million tokens (median across 11 API providers); 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 $0.688 per million tokens for Qwen3.5 35B-A3B versus $0.95 for Qwen3 235B-A22B Thinking 2507 (1.4× as much) and $1.00 for Kimi K2 Instruct (1.5× as much).
Which scores higher on benchmarks?
Qwen3.5 35B-A3B scores higher on the Capabilities Index (ECI): Qwen3.5 35B-A3B 144.1 (#96 of 158), Qwen3 235B-A22B Thinking 2507 143.3 (#100 of 158) and Kimi K2 Instruct 141.3 (#108 of 158). The confidence ranges of the top two overlap (141.6–146.7 vs 141.4–145.1), so treat the gap as small.
Which is better for coding?
There are no published SWE-bench Verified results for Kimi K2 Instruct, Qwen3 235B-A22B Thinking 2507 and Qwen3.5 35B-A3B yet, so there is no like-for-like coding score. On overall capability, Qwen3.5 35B-A3B 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?
Qwen3 235B-A22B Thinking 2507 and Qwen3.5 35B-A3B have the largest context windows (262,144 and 262,144 tokens), against 131,072 for Kimi K2 Instruct. Maximum output per response: Kimi K2 Instruct up to 32,768, Qwen3 235B-A22B Thinking 2507 up to 81,920, Qwen3.5 35B-A3B up to 65,536 tokens.
Which can read images, PDFs, audio or video?
Kimi K2 Instruct accepts text; Qwen3 235B-A22B Thinking 2507 accepts text; Qwen3.5 35B-A3B accepts text, images and video. Qwen3.5 35B-A3B handles the widest range of inputs.
Are any of these open source?
Yes, all three publish their weights (Modified MIT and Apache 2.0), so you can self-host them.
Which is newer?
Qwen3.5 35B-A3B is the newest, released Feb 23, 2026. Qwen3 235B-A22B Thinking 2507 came out Jul 25, 2025; Kimi K2 Instruct came out Jul 11, 2025.
How do you decide the winner?
This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, 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.