GPT-5.6 Terra vs Kimi K3 vs Pareto
Too close to call on our weighted score (GPT-5.6 Terra 45, Kimi K3 42, Pareto 33). The right pick depends on what you value most.
OpenAIGPT-5.6 Terra
45/100- ECI160.1
- Price$2.00 / $12.00
- Context1.05M
Moonshot AIKimi K3
42/100- ECI158.3
- Price$3.00 / $15.00
- Context1.05M
Unbiased
Pareto
33/100- ECI—
- Price$2.57 / $7.72
- Context262K
Too close to call
It is close. Our weighted score puts them within 3 points (GPT-5.6 Terra 45/100, Kimi K3 42/100, Pareto 33/100), so choose by what matters most for your work: Pareto on price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
- CapabilityNot enough dataNo independent benchmark covers every model here yet
- Lowest priceParetoPareto $3.86 · GPT-5.6 Terra $4.50 · Kimi K3 $6.00 per 1M tokens (3:1 blend)
- Longest contextGPT-5.6 Terra and Kimi K3GPT-5.6 Terra 1,050,000 · Kimi K3 1,048,576 · Pareto 262,144 tokens
- Widest inputsGPT-5.6 Terra and Kimi K3GPT-5.6 Terra: Text, Images, PDFs · Kimi K3: Text, Images, Video · Pareto: Text, Images
- Self-hostingKimi K3Publishes downloadable weights
| Measure | Weight | GPT-5.6 Terra | Kimi K3 | Pareto |
|---|---|---|---|---|
| Price | 50% | 18 | 12 | 21 |
| Inputs & features | 30% | 80 | 80 | 50 |
| Context window | 20% | 61 | 61 | 37 |
| Overall | 100% | 45/100 | 42/100 | 33/100 |
Left out because at least one model lacks the data: capability. The remaining weights were rescaled.
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | GPT-5.6 TerraOpenAI | Kimi K3Moonshot AI | ParetoUnbiased |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 160.1 (best) | 158.3 | — |
| ECI rank | #15 of 165 (best) | #20 of 165 | — |
| GPQA DiamondGraduate-level science questions | 93.3% (best) | 93.1% | — |
| FrontierMath Tiers 1–3Research-level mathematics | 86.0% (best) | 72.2% | — |
| OTIS Mock AIME 2024–2025Competition mathematics | 99.7% (best) | 97.2% | — |
| SimpleQA VerifiedShort factual questions | 43.2% | 50.6% (best) | — |
| Price per million tokens | |||
| Input | $2.00 (best) | $3.00 | $2.57 |
| Output | $12.00 | $15.00 | $7.72 (best) |
| Cached input | $0.20 (best) | $0.30 | — |
| Blended (3:1) | $4.50 | $6.00 | $3.86 (best) |
| Long-context rate | Over 272K: $4.00 / $18.00 | Same rate | Same rate |
| Price source | Official OpenAI API | Official Moonshot AI API | Median of 2 providers |
| Limits | |||
| Context window | 1,050,000 tokens (best) | 1,048,576 tokens | 262,144 tokens |
| Max output | 128,000 tokens | 131,072 tokens (best) | 131,072 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | No | No |
| Audio | No | No | No |
| Video | No | Yes | No |
| Reasoning | Yeslow · medium · high · xhigh · max | Yeslow · high · max | No |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | No |
| Availability | |||
| Weights | Proprietary | Open | Proprietary |
| API model ID | gpt-5.6-terra | kimi-k3 | — |
| API providers | 45 | 76 (best) | 2 |
| Released | Jul 9, 2026 | Jul 16, 2026 | Sep 17, 2026 |
| Knowledge cutoff | Feb 16, 2026 | — | — |
What would a month cost?
Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.
- GPT-5.6 Terra$44.00
- Kimi K3$60.00
- Pareto$41.16
Which should you choose?
Which is better: GPT-5.6 Terra, Kimi K3 or Pareto?
It is close. Our weighted score puts them within 3 points (GPT-5.6 Terra 45/100, Kimi K3 42/100, Pareto 33/100), so choose by what matters most for your work: Pareto on price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.
Which is cheaper, GPT-5.6 Terra, Kimi K3 or Pareto?
Pareto is cheaper at $2.57 input / $7.72 output per million tokens (median across 2 API providers). GPT-5.6 Terra costs $2.00 input / $12.00 output per million tokens (official OpenAI API price); Kimi K3 costs $3.00 input / $15.00 output per million tokens (official Moonshot AI API price). At a typical mix of three input tokens to one output token, that is $3.86 per million tokens for Pareto versus $4.50 for GPT-5.6 Terra (1.2× as much) and $6.00 for Kimi K3 (1.6× as much).
Which scores higher on benchmarks?
There is no independent benchmark that covers all three models yet. GPT-5.6 Terra has an ECI of 160.1, Kimi K3 has an ECI of 158.3 and Pareto has not been scored yet.
Which is better for coding?
There are no published SWE-bench Verified results for GPT-5.6 Terra, Kimi K3 and Pareto yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.
Which has the bigger context window?
GPT-5.6 Terra and Kimi K3 have the largest context windows (1,050,000 and 1,048,576 tokens), against 262,144 for Pareto. Maximum output per response: GPT-5.6 Terra up to 128,000, Kimi K3 up to 131,072, Pareto up to 131,072 tokens.
Which can read images, PDFs, audio or video?
GPT-5.6 Terra accepts text, images and PDFs; Kimi K3 accepts text, images and video; Pareto accepts text and images. GPT-5.6 Terra handles the widest range of inputs.
Are any of these open source?
Kimi K3 publishes its weights and can be self-hosted; GPT-5.6 Terra and Pareto is proprietary.
Which is newer?
Pareto is the newest, released Sep 17, 2026. Kimi K3 came out Jul 16, 2026; GPT-5.6 Terra came out Jul 9, 2026. Knowledge cutoff: GPT-5.6 Terra Feb 16, 2026.
How do you decide the winner?
Each model gets a 0–100 score on 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 2, 2026.
