Claude Fable 5.2 and DeepSeek V4 Pro 0423 compared
Claude Fable 5.2, DeepSeek V4 Pro 0423, compared side by side on price, context, and capabilities. All are available on Hormouz through one OpenAI-compatible API.
Anthropic Claude Fable 5.2 | DeepSeek DeepSeek V4 Pro 0423 | |
|---|---|---|
| At a glance | ||
| Context window | 1,000,000 | 1,048,576 |
| Input type | text, image, file | text |
| Output type | text | text |
| Type | text, image | text |
| Pricing | ||
| Input | $4.8/M | $0.955/M |
| Output | $24/M | $1.91/M |
| Cached input | $0.2/M | $0.08/M |
| Cache write | $10/M | — |
| Web search (per 1K calls) | $10/1K | — |
| Persian text | ||
| Tokens per 100 Persian words | 368 | — |
| Persian vs English tokens (same text) | ×2.4 | — |
| Input cost of 1,000 Persian words | $0.018 | — |
| Features | ||
| Thinking (reasoning) | Always on | Can be switched on/off |
| Max output tokens | 128K | 393K |
| Tool use | Yes | Yes |
| Structured output (JSON) | Yes | Yes |
| Prompt caching | Yes | Yes |
| Capabilities | ||
| Capabilities | Vision · Tools · Streaming · Long context | Tools · Streaming · Long context |
| Benchmarks | ||
| Intelligence Index | — | 20 |
| HLEHumanity's Last Exam | — | 11% |
| AA-LCRLong-context reasoning | — | 51% |
| CritPtResearch-level physics reasoning | — | 0% |
| GDPvalEconomically valuable tasks | — | 28% |
| AA OmniscienceBroad factual knowledge accuracy | — | 41% |
| Non-HallucinationRate of avoiding fabricated answers | — | 8% |
| SciCodePython for scientific computing | — | 40% |
Claude Fable 5.2 and DeepSeek V4 Pro 0423: side-by-side summary
Claude Fable 5.2, from Anthropic, has a 1,000,000-token context window and costs $4.8/M for input and $24/M for output on Hormouz (M = one million tokens).
DeepSeek V4 Pro 0423, from DeepSeek, has a 1,048,576-token context window and costs $0.955/M for input and $1.91/M for output on Hormouz (M = one million tokens).
Both are available through one OpenAI-compatible API and one Hormouz key, and you switch between them by changing the model id in your request.
Where they rank
The compared models' scores beside a few models of similar standing.
Intelligence Index
- Claude Fable 5.264.6
- Muse Wonder43.6
- Claude Opus 4.841.8
- GLM 5.3 Flash41.8
- Ling 3.1 Flash41.1
Data from Artificial Analysis; scores out of 100.
FAQ
- Which has the longer context window, Claude Fable 5.2 or DeepSeek V4 Pro 0423?
- DeepSeek V4 Pro 0423 has the longest context window at 1,048,576 tokens, compared with 1,000,000 for Claude Fable 5.2.
- Which is cheaper, Claude Fable 5.2 or DeepSeek V4 Pro 0423?
- DeepSeek V4 Pro 0423 has the lower input price. Claude Fable 5.2: $4.8/M input, $24/M output and DeepSeek V4 Pro 0423: $0.955/M input, $1.91/M output (per million tokens, on Hormouz).
- Which scores higher on intelligence, Claude Fable 5.2 or DeepSeek V4 Pro 0423?
- Claude Fable 5.2 scores higher on the Artificial Analysis Intelligence Index: 64.6 for Claude Fable 5.2 and 20.4 for DeepSeek V4 Pro 0423.
- Do Claude Fable 5.2 and DeepSeek V4 Pro 0423 reason before answering?
- Claude Fable 5.2 always reasons before answering; the reasoning effort setting controls how much. DeepSeek V4 Pro 0423 can have reasoning switched on or off.
- Can I use all 2 of these models through one API?
- Yes. All of them are available on Hormouz through a single OpenAI-compatible API and one API key. Switch between them by changing the model id.
Updated 2026-11-11