Kimi K2 Instruct vs o3 vs Qwen3 235B-A22B Thinking 2507
o3 comes out ahead, 71 to 60 and 55 on our weighted score, though Qwen3 235B-A22B Thinking 2507 is 3.7× cheaper per token.
Price in rating: Excluded · Include price
Moonshot AI
Kimi K2 Instruct
55/100- ECI141.3
- Price$0.57 / $2.30
- Context131K
- Our pick
OpenAI
o3
71/100- ECI147.1
- Price$2.00 / $8.00
- Context200K
Alibaba (Qwen)
Qwen3 235B-A22B Thinking 2507
60/100- ECI143.3
- Price$0.30 / $2.90
- Context262K
o3 is our pick
o3 is the better all-round choice, scoring 71/100 against Qwen3 235B-A22B Thinking 2507 (60) and Kimi K2 Instruct (55). It leads on capability and inputs & features. Qwen3 235B-A22B Thinking 2507 wins on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
- Capabilityo3Capabilities Index (ECI): o3 147.1 · Qwen3 235B-A22B Thinking 2507 143.3 · Kimi K2 Instruct 141.3
- Lowest priceQwen3 235B-A22B Thinking 2507Qwen3 235B-A22B Thinking 2507 $0.95 · Kimi K2 Instruct $1.00 · o3 $3.50 per 1M tokens (3:1 blend)
- Longest contextQwen3 235B-A22B Thinking 2507Qwen3 235B-A22B Thinking 2507 262,144 · o3 200,000 · Kimi K2 Instruct 131,072 tokens
- Widest inputso3Kimi K2 Instruct: Text · o3: Text, Images, PDFs · Qwen3 235B-A22B Thinking 2507: Text
- Self-hostingKimi K2 Instruct and Qwen3 235B-A22B Thinking 2507Publishes downloadable weights (Modified MIT and Apache 2.0)
| Measure | Weight | Kimi K2 Instruct | o3 | Qwen3 235B-A22B Thinking 2507 |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 69 | 75 | 71 |
| Inputs & features | 20% | 25 | 80 | 35 |
| Context window | 13% | 32 | 39 | 43 |
| Overall | 100% | 55/100 | 71/100 | 60/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 141.3 | 147.1 (best) | 143.3 |
| ECI rank | #108 of 158 | #81 of 158 (best) | #100 of 158 |
| GPQA DiamondGraduate-level science questions | — | 81.8% (best) | 80.1% |
| FrontierMath Tiers 1–3Research-level mathematics | — | 33.3% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | — | 84.4% | 86.7% (best) |
| SWE-bench VerifiedFixing real GitHub issues | — | 62.3% | — |
| SimpleQA VerifiedShort factual questions | — | 49.4% (best) | 40.4% |
| Price per million tokens | |||
| Input | $0.57 | $2.00 | $0.30 (best) |
| Output | $2.30 (best) | $8.00 | $2.90 |
| Cached input | — | $0.50 | — |
| Blended (3:1) | $1.00 | $3.50 | $0.95 (best) |
| Long-context rate | Same rate | Same rate | Same rate |
| Price source | Median of 11 providers | Official OpenAI API | Median of 11 providers |
| Limits | |||
| Context window | 131,072 tokens | 200,000 tokens | 262,144 tokens (best) |
| Max output | 32,768 tokens | 100,000 tokens (best) | 81,920 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | No | Yes | No |
| PDFs | No | Yes | No |
| Audio | No | No | No |
| Video | No | No | No |
| Reasoning | No | Yeslow · medium · high | Yes |
| Tool calling | Yes | Yes | Yes |
| Structured output | No | Yes | No |
| Availability | |||
| Weights | OpenModified MIT | Proprietary | OpenApache 2.0 |
| API model ID | — | o3 | — |
| API providers | 11 | 23 (best) | 12 |
| Released | Jul 11, 2025 | Apr 16, 2025 | Jul 25, 2025 |
| Knowledge cutoff | — | May 2024 | — |
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
o3$36.00
Qwen3 235B-A22B Thinking 2507$8.80
Which should you choose?
Which is better: Kimi K2 Instruct, o3 or Qwen3 235B-A22B Thinking 2507?
o3 is the better all-round choice, scoring 71/100 against Qwen3 235B-A22B Thinking 2507 (60) and Kimi K2 Instruct (55). It leads on capability and inputs & features. Qwen3 235B-A22B Thinking 2507 wins on context window. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, Kimi K2 Instruct, o3 or Qwen3 235B-A22B Thinking 2507?
Qwen3 235B-A22B Thinking 2507 is cheaper at $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); o3 costs $2.00 input / $8.00 output per million tokens (official OpenAI API price). At a typical mix of three input tokens to one output token, that is $0.95 per million tokens for Qwen3 235B-A22B Thinking 2507 versus $1.00 for Kimi K2 Instruct (1.1× as much) and $3.50 for o3 (3.7× as much).
Which scores higher on benchmarks?
o3 scores higher on the Capabilities Index (ECI): o3 147.1 (#81 of 158), Qwen3 235B-A22B Thinking 2507 143.3 (#100 of 158) and Kimi K2 Instruct 141.3 (#108 of 158). Their confidence ranges do not overlap (145.6–148.6 vs 141.4–145.1), so the gap is a real one.
Which is better for coding?
There are no published SWE-bench Verified results for Kimi K2 Instruct and Qwen3 235B-A22B Thinking 2507 yet, so there is no like-for-like coding score. On overall capability, o3 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 has the largest context window at 262,144 tokens, against 200,000 for o3 and 131,072 for Kimi K2 Instruct. Maximum output per response: Kimi K2 Instruct up to 32,768, o3 up to 100,000, Qwen3 235B-A22B Thinking 2507 up to 81,920 tokens.
Which can read images, PDFs, audio or video?
Kimi K2 Instruct accepts text; o3 accepts text, images and PDFs; Qwen3 235B-A22B Thinking 2507 accepts text. o3 handles the widest range of inputs.
Are any of these open source?
Kimi K2 Instruct and Qwen3 235B-A22B Thinking 2507 publishes its weights (Modified MIT and Apache 2.0) and can be self-hosted; o3 is proprietary.
Which is newer?
Qwen3 235B-A22B Thinking 2507 is the newest, released Jul 25, 2025. Kimi K2 Instruct came out Jul 11, 2025; o3 came out Apr 16, 2025. Knowledge cutoff: o3 May 2024.
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.