Claude Fable 5.1 vs Gemini 4 Neon vs GPT-6 Astra
Gemini 4 Neon comes out ahead, 69 to 63 and 62 on our weighted score, and it is the cheaper option too.Gemini 4 Neon comes out ahead, 88 to 84 and 83 on our weighted score, and it is the cheaper option too.
Anthropic
Claude Fable 5.1
62/100- ECI167.1
- Price$10.00 / $50.00
- Context1M
- Our pick
Google
Gemini 4 Neon
69/100- ECI—
- Price$3.00 / $15.00
- Context1.05M
OpenAI
GPT-6 Astra
63/100- ECI169.2
- Price$10.00 / $50.00
- Context1.05M
Anthropic
Claude Fable 5.1
83/100- ECI167.1
- Price$10.00 / $50.00
- Context1M
- Our pick
Google
Gemini 4 Neon
88/100- ECI—
- Price$3.00 / $15.00
- Context1.05M
OpenAI
GPT-6 Astra
84/100- ECI169.2
- Price$10.00 / $50.00
- Context1.05M
Gemini 4 Neon is our pick
Gemini 4 Neon is the better all-round choice, scoring 69/100 against GPT-6 Astra (63) and Claude Fable 5.1 (62). It leads on price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Neon has no Capabilities Index score yet.Gemini 4 Neon is the better all-round choice, scoring 88/100 against GPT-6 Astra (84) and Claude Fable 5.1 (83). It leads on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Neon has no Capabilities Index score yet.
- CapabilityGPT-6 AstraShared benchmarks: GPT-6 Astra 89.8% · Gemini 4 Neon 88.8% · Claude Fable 5.1 87.0%
- Lowest priceGemini 4 NeonGemini 4 Neon $6.00 · Claude Fable 5.1 $20.00 · GPT-6 Astra $20.00 per 1M tokens (3:1 blend)
- Longest contextGPT-6 Astra and Gemini 4 NeonGPT-6 Astra 1,050,000 · Gemini 4 Neon 1,048,576 · Claude Fable 5.1 1,000,000 tokens
- Widest inputsGemini 4 NeonClaude Fable 5.1: Text, Images, PDFs · Gemini 4 Neon: Text, Images, PDFs, Audio, Video · GPT-6 Astra: Text, Images, PDFs
- Self-hostingNo confirmed open weightsAll three are available only through APIs
| Measure | Weight | Claude Fable 5.1 | Gemini 4 Neon | GPT-6 Astra |
|---|---|---|---|---|
| CapabilityShared benchmarks | 50% | 87 | 89 | 90 |
| Price | 25% | 0 | 12 | 0 |
| Inputs & features | 15% | 80 | 100 | 80 |
| Context window | 10% | 64 | 65 | 65 |
| Overall | 100% | 62/100 | 69/100 | 63/100 |
| Measure | Weight | Claude Fable 5.1 | Gemini 4 Neon | GPT-6 Astra |
|---|---|---|---|---|
| CapabilityShared benchmarks | 67% | 87 | 89 | 90 |
| Inputs & features | 20% | 80 | 100 | 80 |
| Context window | 13% | 64 | 65 | 65 |
| Overall | 100% | 83/100 | 88/100 | 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 | |||
|---|---|---|---|
| Task | Not published | Multimodal reasoning | Not published |
| Published capabilities | Claude model for demanding reasoning and long-horizon agentic work | Reasons across documents, code, images, audio and video in one context. | GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation. |
| Access | Not published | Paid Gemini API, Google AI Studio, and eligible Gemini subscriptions | Not published |
| Sources | Not published | Publisher documentation · Publisher documentation | Not published |
| Capability | |||
| Capabilities Index (ECI) | 167.1 | Not published | 169.2 (best) |
| ECI rank | #8 of 159 | Not published | #4 of 159 (best) |
| ECI confidence interval | 164.4–170.1 | Not published | 166.4–172.5 |
| ECI evaluation date | Sep 1, 2026 | Not published | Sep 3, 2026 |
| Documented capabilities | Claude model for demanding reasoning and long-horizon agentic work | Reasons across documents, code, images, audio and video in one context. | GPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation. |
| Model parameters | Not published | Not published | Not published |
| Output modalities | Text | Text | Text |
| GPQA DiamondGraduate-level science questions | Not published | 96.3%xhigh (best) | 95.8%max |
| FrontierMath Tiers 1–3Research-level mathematics | 90.2%max | 91.2%xhigh | 93.7%max (best) |
| OTIS Mock AIME 2024–2025Competition mathematics | 100%max | 100%xhigh | 100%max |
| SimpleQA VerifiedShort factual questions | 70.8%max | 75.2%xhigh | 75.6%max (best) |
| Publisher-reported benchmarks | |||
| Terminal-Bench-Science 0.1production safeguards with fallback | 52.6scoreReported resultSep 1, 2026 | Not published | Not published |
| Terminal-Bench 4.0production safeguards with fallback | 55.8scoreReported resultSep 1, 2026 | Not published | Not published |
| GDPval-AA 2production safeguards with fallback | 1853EloReported resultSep 1, 2026 | Not published | Not published |
| OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallback | 77.9partial scoreReported resultSep 1, 2026 | Not published | Not published |
| OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallback | 41.7strict success rateReported resultSep 1, 2026 | Not published | Not published |
| Humanity's Last Examwithout tools; production safeguards with fallback | 60.9scoreReported resultSep 1, 2026 | Not published | Not published |
| Humanity's Last Examwith tools; production safeguards with fallback | 65scoreReported resultSep 1, 2026 | Not published | Not published |
| AutomationBenchproduction safeguards with fallback | 31.4pass rateReported resultSep 1, 2026 | Not published | Not published |
| CursorBench 3.2.0max effort; production safeguards with fallback | 73.4scoreReported resultSep 1, 2026 | Not published | Not published |
| Agents' Last ExamPublisher-reported result | Not published | Not published | 59.3Reported resultSep 3, 2026 |
| OSWorld 2.0 · V2-Offline (v2026.08.08)Publisher-reported result | Not published | Not published | 72.6partial scoreReported resultSep 3, 2026 |
| ScreenSpot-Prono tools | Not published | Not published | 92.7accuracyReported resultSep 3, 2026 |
| AutomationBenchPublisher-reported result | Not published | Not published | 41.4success rateReported resultSep 3, 2026 |
| BenchCADwith tools | Not published | Not published | 95.9geometric overlapReported resultSep 3, 2026 |
| BrowseCompPublisher-reported result | Not published | Not published | 91.5accuracyReported resultSep 3, 2026 |
| Terminal-Bench 4.0Publisher-reported result | Not published | Not published | 57.9success rateReported resultSep 3, 2026 |
| DeepSWE 1.1Publisher-reported result | Not published | Not published | 74.1resolve rateReported resultSep 3, 2026 |
| FrontierCode 1.1 · ExtendedPublisher-reported result | Not published | Not published | 64.5scoreReported resultSep 3, 2026 |
| Terminal-Bench Science 0.1Publisher-reported result | Not published | Not published | 64.6success rateReported resultSep 3, 2026 |
| FrontierMath v2 · Tier 4Publisher-reported result | Not published | Not published | 97.6accuracyReported resultSep 3, 2026 |
| GPQA DiamondPublisher-reported result | Not published | Not published | 96accuracyReported resultSep 3, 2026 |
| Humanity's Last Examwith tools | Not published | Not published | 57.2accuracyReported resultSep 3, 2026 |
| ExploitBenchwithout production safeguards | Not published | Not published | 100success rateReported resultSep 3, 2026 |
| SRE-Benchsingle attempt | Not published | Not published | 88success rateReported resultSep 3, 2026 |
| ARC-AGI-3Publisher-reported result | Not published | Not published | 99.9RHAEReported resultSep 3, 2026 |
| Artificial Analysis Intelligence Index 4.1.1max | Not published | Not published | 61index scoreReported resultSep 3, 2026 |
| Artificial Analysis Coding Agent Index 1.4max | Not published | Not published | 67index scoreReported resultSep 3, 2026 |
| Price per million tokens | |||
| Input | $10.00 | $3.00 (best) | $10.00 |
| Output | $50.00 | $15.00 (best) | $50.00 |
| Cached input | $0.25 | $0.15 (best) | $1.00 |
| Blended (3:1) | $20.00 | $6.00 (best) | $20.00 |
| Long-context rate | Same rate | Same rate | Over 272K: $20.00 / $75.00 |
| Price source | Official Anthropic API | Official Google API | Official OpenAI API |
| Limits | |||
| Context window | 1,000,000 tokens | 1,048,576 tokens | 1,050,000 tokens (best) |
| Max input | Not published | Not published | 922,000 tokens |
| Max output | 128,000 tokens (best) | 65,536 tokens | 128,000 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | Yes | Yes |
| Audio | No | Yes | No |
| Video | No | Yes | No |
| Reasoning | Yeslow · medium · high · xhigh · max | Yeslow · medium · high · xhigh | Yeslow · medium · high · xhigh · max |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | Yes |
| File attachments | Yes | Yes | Yes |
| Temperature control | No | Yes | No |
| Reasoning control | effort | effort | effort |
| Availability | |||
| Weights | Proprietary | Proprietary | Proprietary |
| API model ID | claude-fable-5-1 | gemini-4-neon | gpt-6-astra |
| API providers | 51 | 18 | 56 (best) |
| Released | Sep 1, 2026 | Nov 18, 2026 | Sep 4, 2026 |
| Knowledge cutoff | Jun 2026 | Mar 2026 | Apr 30, 2026 |
| Publisher documentation | Publisher documentation | Publisher documentation · Publisher documentation | Publisher documentation |
| Source metadata updated | Sep 1, 2026 | Nov 18, 2026 | Sep 4, 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.
Claude Fable 5.1$200.00
Gemini 4 Neon$60.00
GPT-6 Astra$200.00
Which should you choose?
Which is better: Claude Fable 5.1, Gemini 4 Neon or GPT-6 Astra?
Gemini 4 Neon is the better all-round choice, scoring 69/100 against GPT-6 Astra (63) and Claude Fable 5.1 (62). It leads on price and inputs & features. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Neon has no Capabilities Index score yet.Gemini 4 Neon is the better all-round choice, scoring 88/100 against GPT-6 Astra (84) and Claude Fable 5.1 (83). It leads on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Neon has no Capabilities Index score yet.
Which is cheaper, Claude Fable 5.1, Gemini 4 Neon or GPT-6 Astra?
Gemini 4 Neon is cheaper at $3.00 input / $15.00 output per million tokens (official Google API price). Claude Fable 5.1 costs $10.00 input / $50.00 output per million tokens (official Anthropic API price); GPT-6 Astra costs $10.00 input / $50.00 output per million tokens (official OpenAI API price). At a typical mix of three input tokens to one output token, that is $6.00 per million tokens for Gemini 4 Neon versus $20.00 for Claude Fable 5.1 (3.3× as much) and $20.00 for GPT-6 Astra (3.3× as much).
Which scores higher on benchmarks?
Not every model here has a Capabilities Index score, so we compare the average of shared benchmarks (FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified): GPT-6 Astra 89.8%, Gemini 4 Neon 88.8% and Claude Fable 5.1 87.0%. On individual benchmarks: FrontierMath Tiers 1–3 — GPT-6 Astra 93.7%, Gemini 4 Neon 91.2%, Claude Fable 5.1 90.2%; OTIS Mock AIME 2024–2025 — Claude Fable 5.1 100%, Gemini 4 Neon 100%, GPT-6 Astra 100%; SimpleQA Verified — GPT-6 Astra 75.6%, Gemini 4 Neon 75.2%, Claude Fable 5.1 70.8%.
Which is better for coding?
There are no published SWE-bench Verified results for Claude Fable 5.1, Gemini 4 Neon and GPT-6 Astra yet, so there is no like-for-like coding score. On overall capability, GPT-6 Astra 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-6 Astra and Gemini 4 Neon 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: Claude Fable 5.1 up to 128,000, Gemini 4 Neon up to 65,536, GPT-6 Astra up to 128,000 tokens.
Which can read images, PDFs, audio or video?
Claude Fable 5.1 accepts text, images and PDFs; Gemini 4 Neon accepts text, images, PDFs, audio and video; GPT-6 Astra accepts text, images and PDFs. Gemini 4 Neon handles the widest range of inputs.
Are any of these open source?
No. Claude Fable 5.1, Gemini 4 Neon and GPT-6 Astra are proprietary and only available through APIs and apps.
Which is newer?
Gemini 4 Neon is the newest, released Nov 18, 2026. GPT-6 Astra came out Sep 4, 2026; Claude Fable 5.1 came out Sep 1, 2026. Knowledge cutoff: Claude Fable 5.1 Jun 2026, Gemini 4 Neon Mar 2026, GPT-6 Astra Apr 30, 2026.
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.