GPT-5.4 vs Claude Opus 5.5 vs Kimi K3
Claude Opus 5.5 comes out ahead, 70 to 65 and 65 on our weighted score, though GPT-5.4 is 30% cheaper per token.
OpenAIGPT-5.4
65/100- ECI156.9
- Price$2.50 / $15.00
- Context1.05M
- Our pick
AnthropicClaude Opus 5.5
70/100- ECI167.4
- Price$4.00 / $20.00
- Context1M
Moonshot AIKimi K3
65/100- ECI157.6
- Price$3.00 / $15.00
- Context1.05M
Claude Opus 5.5 is our pick
Claude Opus 5.5 is the better all-round choice, scoring 70/100 against Kimi K3 (65) and GPT-5.4 (65). It leads on capability. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
- CapabilityClaude Opus 5.5Capabilities Index (ECI): Claude Opus 5.5 167.4 · Kimi K3 157.6 · GPT-5.4 156.9
- Lowest priceGPT-5.4GPT-5.4 $5.63 · Kimi K3 $6.00 · Claude Opus 5.5 $8.00 per 1M tokens (3:1 blend)
- Longest contextGPT-5.4 and Kimi K3GPT-5.4 1,050,000 · Kimi K3 1,048,576 · Claude Opus 5.5 1,000,000 tokens
- Widest inputsSame inputsGPT-5.4: Text, Images, PDFs · Claude Opus 5.5: Text, Images, PDFs · Kimi K3: Text, Images, Video
- Self-hostingKimi K3Publishes downloadable weights
| Measure | Weight | GPT-5.4 | Claude Opus 5.5 | Kimi K3 |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 87 | 100 | 88 |
| Price | 25% | 14 | 7 | 13 |
| Inputs & features | 15% | 80 | 80 | 80 |
| Context window | 10% | 61 | 60 | 61 |
| Overall | 100% | 65/100 | 70/100 | 65/100 |
Every spec in one table
Highlighted cells lead their row. Dashes mean the data is not published.
| Specification | GPT-5.4OpenAI | Claude Opus 5.5Anthropic | Kimi K3Moonshot AI |
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 156.9 | 167.4 (best) | 157.6 |
| ECI rank | #16 of 148 | #1 of 148 (best) | #13 of 148 |
| GPQA DiamondGraduate-level science questions | 93.3% (best) | 90.6% | 93.1% |
| FrontierMath Tiers 1–3Research-level mathematics | 78.6% | 91.2% (best) | 72.2% |
| OTIS Mock AIME 2024–2025Competition mathematics | 97.8% | 100% (best) | 97.2% |
| SWE-bench VerifiedFixing real GitHub issues | 76.9% | — | — |
| SimpleQA VerifiedShort factual questions | 45.1% | 72.2% (best) | 50.6% |
| Price per million tokens | |||
| Input | $2.50 (best) | $4.00 | $3.00 |
| Output | $15.00 (best) | $20.00 | $15.00 (best) |
| Cached input | $0.25 | $0.20 (best) | $0.30 |
| Blended (3:1) | $5.63 (best) | $8.00 | $6.00 |
| Long-context rate | Over 272K: $5.00 / $22.50 | Same rate | Same rate |
| Price source | Official OpenAI API | Official Anthropic API | Official Moonshot AI API |
| Limits | |||
| Context window | 1,050,000 tokens (best) | 1,000,000 tokens | 1,048,576 tokens |
| Max output | 128,000 tokens | 128,000 tokens | 131,072 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | Yes | No |
| Audio | No | No | No |
| Video | No | No | Yes |
| Reasoning | Yeslow · medium · high · xhigh | Yeslow · medium · high · xhigh · max | Yeslow · high · max |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | Yes |
| Availability | |||
| Weights | Proprietary | Proprietary | Open |
| API model ID | gpt-5.4 | claude-opus-5-5 | kimi-k3 |
| API providers | 36 | 28 | 68 (best) |
| Released | Mar 5, 2026 | Sep 22, 2026 | Jul 16, 2026 |
| Knowledge cutoff | Aug 31, 2025 | Jun 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-5.4$55.00
- Claude Opus 5.5$80.00
- Kimi K3$60.00
Which should you choose?
Which is better: GPT-5.4, Claude Opus 5.5 or Kimi K3?
Claude Opus 5.5 is the better all-round choice, scoring 70/100 against Kimi K3 (65) and GPT-5.4 (65). It leads on capability. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.
Which is cheaper, GPT-5.4, Claude Opus 5.5 or Kimi K3?
GPT-5.4 is cheaper at $2.50 input / $15.00 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); Claude Opus 5.5 costs $4.00 input / $20.00 output per million tokens (official Anthropic API price). At a typical mix of three input tokens to one output token, that is $5.63 per million tokens for GPT-5.4 versus $6.00 for Kimi K3 (1.1× as much) and $8.00 for Claude Opus 5.5 (1.4× as much).
Which scores higher on benchmarks?
Claude Opus 5.5 scores higher on the Capabilities Index (ECI): Claude Opus 5.5 167.4 (#1 of 148), Kimi K3 157.6 (#13 of 148) and GPT-5.4 156.9 (#16 of 148). Their confidence ranges do not overlap (164.0–172.0 vs 154.9–160.4), so the gap is a real one. On individual benchmarks: GPQA Diamond — GPT-5.4 93.3%, Kimi K3 93.1%, Claude Opus 5.5 90.6%; FrontierMath Tiers 1–3 — Claude Opus 5.5 91.2%, GPT-5.4 78.6%, Kimi K3 72.2%; OTIS Mock AIME 2024–2025 — Claude Opus 5.5 100%, GPT-5.4 97.8%, Kimi K3 97.2%; SimpleQA Verified — Claude Opus 5.5 72.2%, Kimi K3 50.6%, GPT-5.4 45.1%.
Which is better for coding?
There are no published SWE-bench Verified results for Claude Opus 5.5 and Kimi K3 yet, so there is no like-for-like coding score. On overall capability, Claude Opus 5.5 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?
GPT-5.4 and Kimi K3 have the largest context windows (1,050,000 and 1,048,576 tokens), against 1,000,000 for Claude Opus 5.5. Maximum output per response: GPT-5.4 up to 128,000, Claude Opus 5.5 up to 128,000, Kimi K3 up to 131,072 tokens.
Which can read images, PDFs, audio or video?
GPT-5.4 accepts text, images and PDFs; Claude Opus 5.5 accepts text, images and PDFs; Kimi K3 accepts text, images and video. They handle the same number of input types.
Are any of these open source?
Kimi K3 publishes its weights and can be self-hosted; GPT-5.4 and Claude Opus 5.5 is proprietary.
Which is newer?
Claude Opus 5.5 is the newest, released Sep 22, 2026. Kimi K3 came out Jul 16, 2026; GPT-5.4 came out Mar 5, 2026. Knowledge cutoff: GPT-5.4 Aug 31, 2025, Claude Opus 5.5 Jun 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 Oct 4, 2026.
