GPT-5.4 vs Claude Fable 5.1 vs Gemini 3.1 Pro Preview
Too close to call on our weighted score (Gemini 3.1 Pro Preview 68, Claude Fable 5.1 65, GPT-5.4 65). The right pick depends on what you value most.Too close to call on our weighted score (Claude Fable 5.1 87, Gemini 3.1 Pro Preview 84, GPT-5.4 82). The right pick depends on what you value most.
OpenAI
GPT-5.4
65/100- ECI157.9
- Price$2.50 / $15.00
- Context1.05M
Anthropic
Claude Fable 5.1
65/100- ECI167.1
- Price$10.00 / $50.00
- Context1M
Google
Gemini 3.1 Pro Preview
68/100- ECI156.3
- Price$2.00 / $12.00
- Context1.05M
OpenAI
GPT-5.4
82/100- ECI157.9
- Price$2.50 / $15.00
- Context1.05M
Anthropic
Claude Fable 5.1
87/100- ECI167.1
- Price$10.00 / $50.00
- Context1M
Google
Gemini 3.1 Pro Preview
84/100- ECI156.3
- Price$2.00 / $12.00
- Context1.05M
Too close to call
It is close. Our weighted score puts them within 2 points (Gemini 3.1 Pro Preview 68/100, Claude Fable 5.1 65/100, GPT-5.4 65/100), so choose by what matters most for your work: Claude Fable 5.1 for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.It is close. Our weighted score puts them within 3 points (Claude Fable 5.1 87/100, Gemini 3.1 Pro Preview 84/100, GPT-5.4 82/100), so choose by what matters most for your work: Claude Fable 5.1 for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 67%, inputs & features 20%, context window 13%.
- CapabilityClaude Fable 5.1Capabilities Index (ECI): Claude Fable 5.1 167.1 · GPT-5.4 157.9 · Gemini 3.1 Pro Preview 156.3
- Lowest priceGemini 3.1 Pro PreviewGemini 3.1 Pro Preview $4.50 · GPT-5.4 $5.63 · Claude Fable 5.1 $20.00 per 1M tokens (3:1 blend)
- Longest contextGPT-5.4 and Gemini 3.1 Pro PreviewGPT-5.4 1,050,000 · Gemini 3.1 Pro Preview 1,048,576 · Claude Fable 5.1 1,000,000 tokens
- Widest inputsGemini 3.1 Pro PreviewGPT-5.4: Text, Images, PDFs · Claude Fable 5.1: Text, Images, PDFs · Gemini 3.1 Pro Preview: Text, Images, PDFs, Audio, Video
- Self-hostingNo confirmed open weightsAll three are available only through APIs
| Measure | Weight | GPT-5.4 | Claude Fable 5.1 | Gemini 3.1 Pro Preview |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 50% | 85 | 94 | 84 |
| Price | 25% | 14 | 0 | 18 |
| Inputs & features | 15% | 80 | 80 | 100 |
| Context window | 10% | 65 | 64 | 65 |
| Overall | 100% | 65/100 | 65/100 | 68/100 |
| Measure | Weight | GPT-5.4 | Claude Fable 5.1 | Gemini 3.1 Pro Preview |
|---|---|---|---|---|
| CapabilityCapabilities Index (ECI) | 67% | 85 | 94 | 84 |
| Inputs & features | 20% | 80 | 80 | 100 |
| Context window | 13% | 65 | 64 | 65 |
| Overall | 100% | 82/100 | 87/100 | 84/100 |
Every spec in one table
Highlighted cells lead their row. Unavailable specifications are marked “Not published”. Publisher results include their evaluation settings and source.
| Specification | |||
|---|---|---|---|
| Capability | |||
| Capabilities Index (ECI) | 157.9 | 167.1 (best) | 156.3 |
| ECI rank | #29 of 159 | #8 of 159 (best) | #41 of 159 |
| ECI confidence interval | 156.0–160.0 | 164.4–170.1 | 154.5–158.2 |
| ECI evaluation date | Mar 5, 2026 | Sep 1, 2026 | Feb 19, 2026 |
| Documented capabilities | Agent-ready GPT for coding and computer-use workflows at a lower cost | Claude model for demanding reasoning and long-horizon agentic work | Reasoning-first Gemini preview for agentic coding and complex problem solving |
| Model parameters | Not published | Not published | Not published |
| Output modalities | Text | Text | Text |
| GPQA DiamondGraduate-level science questions | 93.3%xhigh | Not published | 94.4%high (best) |
| FrontierMath Tiers 1–3Research-level mathematics | 78.6%xhigh | 90.2%max (best) | 59.7% |
| OTIS Mock AIME 2024–2025Competition mathematics | 97.8%high | 100%max (best) | 95.6% |
| SWE-bench VerifiedFixing real GitHub issues | 76.9%high (best) | Not published | 75.6% |
| SimpleQA VerifiedShort factual questions | 45.1%xhigh | 70.8%max | 73.5%high (best) |
| Publisher-reported benchmarks | |||
| SWE-Bench Pro · publicPublisher-reported result | 59.1resolve rateReported result | Not published | 46.1resolve rateReported result |
| SWE-Atlas Codebase QnAPublisher-reported result | 40.8scoreReported result | Not published | 13.5scoreReported result |
| SWE-Atlas RefactoringPublisher-reported result | 44.29scoreReported result | Not published | 33.81scoreReported result |
| SWE-Atlas Test WritingPublisher-reported result | 44.36scoreReported result | Not published | 29.84scoreReported result |
| Artificial Analysis Coding Agent Indexmedium | 53.6average pass@1Reported result | Not published | Not published |
| SWE-Atlas Codebase QnAmedium | 72.4pass@1Reported result | Not published | Not published |
| SWE-Bench Pro · hard-aamedium | 18.4pass@1Reported result | Not published | Not published |
| Terminal-Bench 2.1medium | 69.8pass@1Reported result | Not published | Not published |
| Terminal-Bench 2.0Publisher-reported result | 75.1success rateReported resultApr 23, 2026 | Not published | Not published |
| GPQA DiamondPublisher-reported result | 92.8accuracyReported resultApr 23, 2026 | Not published | 94.3accuracyReported resultApr 23, 2026 |
| Humanity's Last Examno tools | 39.8accuracyReported resultApr 23, 2026 | Not published | Not published |
| Humanity's Last Examwith tools | 52.1accuracyReported resultApr 23, 2026 | Not published | Not published |
| OSWorld-VerifiedPublisher-reported result | 75success rateReported resultApr 23, 2026 | Not published | 76.2success rateReported resultMay 19, 2026 |
| BrowseCompPublisher-reported result | 82.7accuracyReported resultApr 23, 2026 | Not published | Not published |
| GDPvalPublisher-reported result | 83wins or tiesReported resultApr 23, 2026 | Not published | Not published |
| ARC-AGI-2Verified | 73.3accuracyReported resultApr 23, 2026 | Not published | Not published |
| FrontierMath · Tier 1-3Publisher-reported result | 47.6accuracyReported resultApr 23, 2026 | Not published | Not published |
| FrontierMath · Tier 4Publisher-reported result | 27.1accuracyReported resultApr 23, 2026 | Not published | Not published |
| MMMU Prono tools | 81.2accuracyReported resultApr 23, 2026 | Not published | 80.5accuracyReported resultMay 19, 2026 |
| Terminal-Bench-Science 0.1production safeguards with fallback | Not published | 52.6scoreReported resultSep 1, 2026 | Not published |
| Terminal-Bench 4.0production safeguards with fallback | Not published | 55.8scoreReported resultSep 1, 2026 | Not published |
| GDPval-AA 2production safeguards with fallback | Not published | 1853EloReported resultSep 1, 2026 | Not published |
| OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallback | Not published | 77.9partial scoreReported resultSep 1, 2026 | Not published |
| OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallback | Not published | 41.7strict success rateReported resultSep 1, 2026 | Not published |
| Humanity's Last Examwithout tools; production safeguards with fallback | Not published | 60.9scoreReported resultSep 1, 2026 | Not published |
| Humanity's Last Examwith tools; production safeguards with fallback | Not published | 65scoreReported resultSep 1, 2026 | Not published |
| AutomationBenchproduction safeguards with fallback | Not published | 31.4pass rateReported resultSep 1, 2026 | Not published |
| CursorBench 3.2.0max effort; production safeguards with fallback | Not published | 73.4scoreReported resultSep 1, 2026 | Not published |
| SWE-Bench ProPublisher-reported result | Not published | Not published | 54.2resolve rateReported resultMay 28, 2026 |
| Terminal-Bench 2.1Publisher-reported result | Not published | Not published | 70.3success rateReported resultMay 28, 2026 |
| Artificial Analysis Coding Agent Indexhigh | Not published | Not published | 43average pass@1Reported result |
| SWE-Atlas Codebase QnAhigh | Not published | Not published | 45.6pass@1Reported result |
| SWE-Bench Pro · hard-aahigh | Not published | Not published | 15.1pass@1Reported result |
| Terminal-Bench 2.1high | Not published | Not published | 68.3pass@1Reported result |
| Humanity's Last Exam · full set, text + MMPublisher-reported result | Not published | Not published | 44.4accuracyReported resultMay 19, 2026 |
| ARC-AGI-2Publisher-reported result | Not published | Not published | 77.1accuracyReported resultMay 19, 2026 |
| MCP AtlasPublisher-reported result | Not published | Not published | 78.2success rateReported resultMay 19, 2026 |
| CharXiv Reasoningno tools | Not published | Not published | 83.3accuracyReported resultMay 19, 2026 |
| GDPval-AAPublisher-reported result | Not published | Not published | 1314EloReported resultMay 19, 2026 |
| Price per million tokens | |||
| Input | $2.50 | $10.00 | $2.00 (best) |
| Output | $15.00 | $50.00 | $12.00 (best) |
| Cached input | $0.25 | $0.25 | $0.20 (best) |
| Blended (3:1) | $5.63 | $20.00 | $4.50 (best) |
| Long-context rate | Over 272K: $5.00 / $22.50 | Same rate | Over 200K: $4.00 / $18.00 |
| Price source | Official OpenAI API | Official Anthropic API | Official Google API |
| Limits | |||
| Context window | 1,050,000 tokens (best) | 1,000,000 tokens | 1,048,576 tokens |
| Max input | 922,000 tokens | Not published | Not published |
| Max output | 128,000 tokens (best) | 128,000 tokens (best) | 65,536 tokens |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | Yes | Yes |
| Audio | No | No | Yes |
| Video | No | No | Yes |
| Reasoning | Yeslow · medium · high · xhigh | Yeslow · medium · high · xhigh · max | Yeslow · medium · high |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | Yes |
| File attachments | Yes | Yes | Yes |
| Temperature control | Yes | No | Yes |
| Reasoning control | effort | effort | effort |
| Availability | |||
| Weights | Proprietary | Proprietary | Proprietary |
| API model ID | gpt-5.4 | claude-fable-5-1 | gemini-3.1-pro-preview |
| API providers | 50 | 51 (best) | 40 |
| Released | Mar 5, 2026 | Sep 1, 2026 | Feb 19, 2026 |
| Knowledge cutoff | Aug 31, 2025 | Jun 2026 | Jan 2025 |
| Publisher documentation | Publisher documentation | Publisher documentation | Publisher documentation |
| Source metadata updated | Mar 5, 2026 | Sep 1, 2026 | Feb 19, 2026 |
| Data sources | models.dev · Epoch AI | models.dev · Epoch AI | models.dev · Epoch AI |
| Catalog checked | Nov 22, 2026 | Nov 22, 2026 | Nov 22, 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 Fable 5.1$200.00
Gemini 3.1 Pro Preview$44.00
Which should you choose?
Which is better: GPT-5.4, Claude Fable 5.1 or Gemini 3.1 Pro Preview?
It is close. Our weighted score puts them within 2 points (Gemini 3.1 Pro Preview 68/100, Claude Fable 5.1 65/100, GPT-5.4 65/100), so choose by what matters most for your work: Claude Fable 5.1 for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.It is close. Our weighted score puts them within 3 points (Claude Fable 5.1 87/100, Gemini 3.1 Pro Preview 84/100, GPT-5.4 82/100), so choose by what matters most for your work: Claude Fable 5.1 for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 67%, inputs & features 20%, context window 13%.
Which is cheaper, GPT-5.4, Claude Fable 5.1 or Gemini 3.1 Pro Preview?
Gemini 3.1 Pro Preview is cheaper at $2.00 input / $12.00 output per million tokens (official Google API price). GPT-5.4 costs $2.50 input / $15.00 output per million tokens (official OpenAI API price); Claude Fable 5.1 costs $10.00 input / $50.00 output per million tokens (official Anthropic API price). At a typical mix of three input tokens to one output token, that is $4.50 per million tokens for Gemini 3.1 Pro Preview versus $5.63 for GPT-5.4 (1.3× as much) and $20.00 for Claude Fable 5.1 (4.4× as much).
Which scores higher on benchmarks?
Claude Fable 5.1 scores higher on the Capabilities Index (ECI): Claude Fable 5.1 167.1 (#8 of 159), GPT-5.4 157.9 (#29 of 159) and Gemini 3.1 Pro Preview 156.3 (#41 of 159). Their confidence ranges do not overlap (164.4–170.1 vs 156.0–160.0), so the gap is a real one. On individual benchmarks: FrontierMath Tiers 1–3 — Claude Fable 5.1 90.2%, GPT-5.4 78.6%, Gemini 3.1 Pro Preview 59.7%; OTIS Mock AIME 2024–2025 — Claude Fable 5.1 100%, GPT-5.4 97.8%, Gemini 3.1 Pro Preview 95.6%; SimpleQA Verified — Gemini 3.1 Pro Preview 73.5%, Claude Fable 5.1 70.8%, GPT-5.4 45.1%.
Which is better for coding?
There are no published SWE-bench Verified results for Claude Fable 5.1 yet, so there is no like-for-like coding score. On overall capability, Claude Fable 5.1 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 Gemini 3.1 Pro Preview have the largest context windows (1,050,000 and 1,048,576 tokens), against 1,000,000 for Claude Fable 5.1. Maximum output per response: GPT-5.4 up to 128,000, Claude Fable 5.1 up to 128,000, Gemini 3.1 Pro Preview up to 65,536 tokens.
Which can read images, PDFs, audio or video?
GPT-5.4 accepts text, images and PDFs; Claude Fable 5.1 accepts text, images and PDFs; Gemini 3.1 Pro Preview accepts text, images, PDFs, audio and video. Gemini 3.1 Pro Preview handles the widest range of inputs.
Are any of these open source?
No. GPT-5.4, Claude Fable 5.1 and Gemini 3.1 Pro Preview are proprietary and only available through APIs and apps.
Which is newer?
Claude Fable 5.1 is the newest, released Sep 1, 2026. GPT-5.4 came out Mar 5, 2026; Gemini 3.1 Pro Preview came out Feb 19, 2026. Knowledge cutoff: GPT-5.4 Aug 31, 2025, Claude Fable 5.1 Jun 2026, Gemini 3.1 Pro Preview Jan 2025.
How do you decide the winner?
This rating uses 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 22, 2026.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 22, 2026.