GPT-5.6 Terra vs Kimi K3 vs Muse Wonder
Too close to call on our weighted score (Muse Wonder 68, GPT-5.6 Terra 67, Kimi K3 65). The right pick depends on what you value most.
OpenAIGPT-5.6 Terra
67/100- ECI160.1
- Price$2.00 / $12.00
- Context1.05M
Moonshot AIKimi K3
65/100- ECI158.3
- Price$3.00 / $15.00
- Context1.05M
MetaMuse Wonder
68/100- ECI161.7
- Price$4.00 / $20.00
- Context1.05M
Too close to call
It is close. Our weighted score puts them within 1 points (Muse Wonder 68/100, GPT-5.6 Terra 67/100, Kimi K3 65/100), so choose by what matters most for your work: Muse Wonder for raw capability and GPT-5.6 Terra on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityMuse WonderCapabilities Index (ECI): Muse Wonder 161.7 · GPT-5.6 Terra 160.1 · Kimi K3 158.3
- Lowest priceGPT-5.6 TerraGPT-5.6 Terra $4.50 · Kimi K3 $6.00 · Muse Wonder $8.00 per 1M tokens (3:1 blend)
- Longest contextAbout the sameGPT-5.6 Terra 1,050,000 · Kimi K3 1,048,576 · Muse Wonder 1,048,576 tokens
- Widest inputsMuse WonderGPT-5.6 Terra: Text, Images, PDFs · Kimi K3: Text, Images, Video · Muse Wonder: Text, Images, PDFs, Audio, Video
- Self-hostingKimi K3Publishes downloadable weights
| Measure | Weight | GPT-5.6 Terra | Kimi K3 | Muse Wonder |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 89 | 87 | 91 |
| Price | 25% | 18 | 12 | 7 |
| Inputs & features | 15% | 80 | 80 | 100 |
| Context window | 10% | 61 | 61 | 61 |
| Overall | 100% | 67/100 | 65/100 | 68/100 |
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 | Muse WonderMeta |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 160.1 | 158.3 | 161.7 (best) |
| ECI rank | #15 of 165 | #20 of 165 | #13 of 165 (best) |
| GPQA DiamondGraduate-level science questions | 93.3% (best) | 93.1% | — |
| FrontierMath Tiers 1–3Research-level mathematics | 86.0% (best) | 72.2% | 82.1% |
| OTIS Mock AIME 2024–2025Competition mathematics | 99.7% (best) | 97.2% | 99.4% |
| SimpleQA VerifiedShort factual questions | 43.2% | 50.6% (best) | — |
| Price per million tokens | |||
| Input | $2.00 (best) | $3.00 | $4.00 |
| Output | $12.00 (best) | $15.00 | $20.00 |
| Cached input | $0.20 (best) | $0.30 | $0.20 (best) |
| Blended (3:1) | $4.50 (best) | $6.00 | $8.00 |
| Long-context rate | Over 272K: $4.00 / $18.00 | Same rate | Same rate |
| Price source | Official OpenAI API | Official Moonshot AI API | Official Meta API |
| Limits | |||
| Context window | 1,050,000 tokens (best) | 1,048,576 tokens | 1,048,576 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 | Yes |
| Audio | No | No | Yes |
| Video | No | Yes | Yes |
| Reasoning | Yeslow · medium · high · xhigh · max | Yeslow · high · max | Yesminimal · low · medium · high · xhigh · max |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | Yes |
| Availability | |||
| Weights | Proprietary | Open | Proprietary |
| API model ID | gpt-5.6-terra | kimi-k3 | muse-wonder-1.0 |
| API providers | 45 | 76 (best) | 9 |
| Released | Jul 9, 2026 | Jul 16, 2026 | Oct 12, 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
- Muse Wonder$80.00
Which should you choose?
Which is better: GPT-5.6 Terra, Kimi K3 or Muse Wonder?
It is close. Our weighted score puts them within 1 points (Muse Wonder 68/100, GPT-5.6 Terra 67/100, Kimi K3 65/100), so choose by what matters most for your work: Muse Wonder for raw capability and GPT-5.6 Terra on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, GPT-5.6 Terra, Kimi K3 or Muse Wonder?
GPT-5.6 Terra is cheaper at $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); Muse Wonder costs $4.00 input / $20.00 output per million tokens (official Meta API price). At a typical mix of three input tokens to one output token, that is $4.50 per million tokens for GPT-5.6 Terra versus $6.00 for Kimi K3 (1.3× as much) and $8.00 for Muse Wonder (1.8× as much).
Which scores higher on benchmarks?
Muse Wonder scores higher on the Capabilities Index (ECI): Muse Wonder 161.7 (#13 of 165), GPT-5.6 Terra 160.1 (#15 of 165) and Kimi K3 158.3 (#20 of 165). The confidence ranges of the top two overlap (159.2–164.6 vs 157.8–163.4), so treat the gap as small. On individual benchmarks: FrontierMath Tiers 1–3 — GPT-5.6 Terra 86.0%, Muse Wonder 82.1%, Kimi K3 72.2%; OTIS Mock AIME 2024–2025 — GPT-5.6 Terra 99.7%, Muse Wonder 99.4%, Kimi K3 97.2%.
Which is better for coding?
There are no published SWE-bench Verified results for GPT-5.6 Terra, Kimi K3 and Muse Wonder yet, so there is no like-for-like coding score. On overall capability, Muse Wonder 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?
Their context windows are effectively the same size: GPT-5.6 Terra 1,050,000, Kimi K3 1,048,576 and Muse Wonder 1,048,576 tokens. Maximum output per response: GPT-5.6 Terra up to 128,000, Kimi K3 up to 131,072, Muse Wonder 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; Muse Wonder accepts text, images, PDFs, audio and video. Muse Wonder 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 Muse Wonder is proprietary.
Which is newer?
Muse Wonder is the newest, released Oct 12, 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.
