Gemini 4 Neon vs GPT-5.6 Sol vs Claude Sonnet 5.5
Gemini 4 Neon comes out ahead, 70 to 65 and 64 on our weighted score, though Claude Sonnet 5.5 is 33% cheaper per token.Gemini 4 Neon comes out ahead, 89 to 83 and 80 on our weighted score, though Claude Sonnet 5.5 is 33% cheaper per token.
- Our pick
Google
Gemini 4 Neon
70/100- ECI—
- Price$3.00 / $15.00
- Context1.05M
OpenAI
GPT-5.6 Sol
64/100- ECI163.0
- Price$4.00 / $20.00
- Context1.05M
Anthropic
Claude Sonnet 5.5
65/100- ECI166.6
- Price$2.00 / $10.00
- Context1M
- Our pick
Google
Gemini 4 Neon
89/100- ECI—
- Price$3.00 / $15.00
- Context1.05M
OpenAI
GPT-5.6 Sol
83/100- ECI163.0
- Price$4.00 / $20.00
- Context1.05M
Anthropic
Claude Sonnet 5.5
80/100- ECI166.6
- Price$2.00 / $10.00
- Context1M
Gemini 4 Neon is our pick
Gemini 4 Neon is the better all-round choice, scoring 70/100 against Claude Sonnet 5.5 (65) and GPT-5.6 Sol (64). It leads on capability and inputs & features. Claude Sonnet 5.5 wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (GPQA Diamond, 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 89/100 against GPT-5.6 Sol (83) and Claude Sonnet 5.5 (80). It leads on capability and inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond, FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified), because Gemini 4 Neon has no Capabilities Index score yet.
- CapabilityGemini 4 NeonShared benchmarks: Gemini 4 Neon 90.7% · GPT-5.6 Sol 88.1% · Claude Sonnet 5.5 82.7%
- Lowest priceClaude Sonnet 5.5Claude Sonnet 5.5 $4.00 · Gemini 4 Neon $6.00 · GPT-5.6 Sol $8.00 per 1M tokens (3:1 blend)
- Longest contextGPT-5.6 Sol and Gemini 4 NeonGPT-5.6 Sol 1,050,000 · Gemini 4 Neon 1,048,576 · Claude Sonnet 5.5 1,000,000 tokens
- Widest inputsGemini 4 NeonGemini 4 Neon: Text, Images, PDFs, Audio, Video · GPT-5.6 Sol: Text, Images, PDFs · Claude Sonnet 5.5: Text, Images, PDFs
- Self-hostingNo confirmed open weightsAll three are available only through APIs
| Measure | Weight | Gemini 4 Neon | GPT-5.6 Sol | Claude Sonnet 5.5 |
|---|---|---|---|---|
| CapabilityShared benchmarks | 50% | 91 | 88 | 83 |
| Price | 25% | 12 | 7 | 20 |
| Inputs & features | 15% | 100 | 80 | 80 |
| Context window | 10% | 65 | 65 | 64 |
| Overall | 100% | 70/100 | 64/100 | 65/100 |
| Measure | Weight | Gemini 4 Neon | GPT-5.6 Sol | Claude Sonnet 5.5 |
|---|---|---|---|---|
| CapabilityShared benchmarks | 67% | 91 | 88 | 83 |
| Inputs & features | 20% | 100 | 80 | 80 |
| Context window | 13% | 65 | 65 | 64 |
| Overall | 100% | 89/100 | 83/100 | 80/100 |
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 | Multimodal reasoning | Not published | Not published |
| Published capabilities | Reasons across documents, code, images, audio and video in one context. | Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows | Fast Claude model for everyday coding, agents, and knowledge work |
| Access | Paid Gemini API, Google AI Studio, and eligible Gemini subscriptions | Not published | Not published |
| Sources | Publisher documentation · Publisher documentation | Not published | Not published |
| Capability | |||
| Capabilities Index (ECI) | Not published | 163.0 | 166.6 (best) |
| ECI rank | Not published | #15 of 159 | #10 of 159 (best) |
| ECI confidence interval | Not published | 160.7–165.6 | 163.9–169.6 |
| ECI evaluation date | Not published | Jul 9, 2026 | Sep 28, 2026 |
| Documented capabilities | Reasons across documents, code, images, audio and video in one context. | Frontier GPT-5.6 model for complex professional work, coding, and agentic workflows | Fast Claude model for everyday coding, agents, and knowledge work |
| Model parameters | Not published | Not published | Not published |
| Output modalities | Text | Text | Text |
| GPQA DiamondGraduate-level science questions | 96.3%xhigh (best) | 93.5%max | 95.6%max |
| FrontierMath Tiers 1–3Research-level mathematics | 91.2%xhigh (best) | 89.1%max | 88.8%max |
| OTIS Mock AIME 2024–2025Competition mathematics | 100%xhigh | 100%max | 100%max |
| SimpleQA VerifiedShort factual questions | 75.2%xhigh (best) | 69.7%max | 46.5%max |
| Publisher-reported benchmarks | |||
| SWE-Bench ProPublisher-reported result | Not published | 64.6resolve rateReported resultJul 9, 2026 | Not published |
| Terminal-Bench 2.1Publisher-reported result | Not published | 88.8success rateReported resultJul 9, 2026 | Not published |
| DeepSWE 1.1Publisher-reported result | Not published | 72.7resolve rateReported resultJul 9, 2026 | Not published |
| GPQA DiamondPublisher-reported result | Not published | 94.6accuracyReported resultJul 9, 2026 | Not published |
| FrontierMath v2 · Tier 1-3Publisher-reported result | Not published | 89accuracyReported resultJul 9, 2026 | Not published |
| BrowseCompPublisher-reported result | Not published | 90.4accuracyReported resultJul 9, 2026 | Not published |
| OSWorld 2.0Publisher-reported result | Not published | 62.6success rateReported resultJul 9, 2026 | Not published |
| MMMU Prono tools | Not published | 83accuracyReported resultJul 9, 2026 | Not published |
| Agents' Last ExamPublisher-reported result | Not published | 52.7Reported resultJul 9, 2026 | Not published |
| ToolathlonPublisher-reported result | Not published | 58success rateReported resultJul 9, 2026 | Not published |
| Artificial Analysis Intelligence Index 4.1max | Not published | 58.9index scoreReported resultJul 9, 2026 | Not published |
| Artificial Analysis Coding Agent Index 1.1max | Not published | 80index scoreReported resultJul 9, 2026 | Not published |
| Terminal-Bench 4.0Publisher-reported result | Not published | Not published | 70.6scoreReported resultSep 28, 2026 |
| FrontierCode 1.1 · Mainmax effort | Not published | Not published | 46.2scoreReported resultSep 28, 2026 |
| CursorBench 4.0Publisher-reported result | Not published | Not published | 55.5scoreReported resultSep 28, 2026 |
| GDPval-AA 2.1Publisher-reported result | Not published | Not published | 1844EloReported resultSep 28, 2026 |
| Humanity's Last Examwith tools | Not published | Not published | 64.5scoreReported resultSep 28, 2026 |
| OSWorld 2.1Publisher-reported result | Not published | Not published | 80.1partial scoreReported resultSep 28, 2026 |
| Price per million tokens | |||
| Input | $3.00 | $4.00 | $2.00 (best) |
| Output | $15.00 | $20.00 | $10.00 (best) |
| Cached input | $0.15 | $0.40 | $0.10 (best) |
| Blended (3:1) | $6.00 | $8.00 | $4.00 (best) |
| Long-context rate | Same rate | Over 272K: $8.00 / $30.00 | Same rate |
| Price source | Official Google API | Official OpenAI API | Official Anthropic API |
| Limits | |||
| Context window | 1,048,576 tokens | 1,050,000 tokens (best) | 1,000,000 tokens |
| Max input | Not published | 922,000 tokens | Not published |
| Max output | 65,536 tokens | 128,000 tokens (best) | 128,000 tokens (best) |
| Inputs and features | |||
| Text | Yes | Yes | Yes |
| Images | Yes | Yes | Yes |
| PDFs | Yes | Yes | Yes |
| Audio | Yes | No | No |
| Video | Yes | No | No |
| Reasoning | Yeslow · medium · high · xhigh | Yeslow · medium · high · xhigh · max | Yeslow · medium · high · xhigh · max |
| Tool calling | Yes | Yes | Yes |
| Structured output | Yes | Yes | Yes |
| File attachments | Yes | Yes | Yes |
| Temperature control | Yes | Yes | No |
| Reasoning control | effort | effort | effort |
| Availability | |||
| Weights | Proprietary | Proprietary | Proprietary |
| API model ID | gemini-4-neon | gpt-5.6-sol | claude-sonnet-5-5 |
| API providers | 18 | 61 (best) | 57 |
| Released | Nov 18, 2026 | Jul 9, 2026 | Sep 28, 2026 |
| Knowledge cutoff | Mar 2026 | Feb 16, 2026 | Jun 2026 |
| Publisher documentation | Publisher documentation · Publisher documentation | Publisher documentation | Publisher documentation |
| Source metadata updated | Nov 18, 2026 | Jul 9, 2026 | Sep 28, 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.
Gemini 4 Neon$60.00
GPT-5.6 Sol$80.00
Claude Sonnet 5.5$40.00
Which should you choose?
Which is better: Gemini 4 Neon, GPT-5.6 Sol or Claude Sonnet 5.5?
Gemini 4 Neon is the better all-round choice, scoring 70/100 against Claude Sonnet 5.5 (65) and GPT-5.6 Sol (64). It leads on capability and inputs & features. Claude Sonnet 5.5 wins on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%. Capability uses the average of shared benchmarks (GPQA Diamond, 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 89/100 against GPT-5.6 Sol (83) and Claude Sonnet 5.5 (80). It leads on capability and inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%. Capability uses the average of shared benchmarks (GPQA Diamond, 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, Gemini 4 Neon, GPT-5.6 Sol or Claude Sonnet 5.5?
Claude Sonnet 5.5 is cheaper at $2.00 input / $10.00 output per million tokens (official Anthropic API price). Gemini 4 Neon costs $3.00 input / $15.00 output per million tokens (official Google API price); GPT-5.6 Sol costs $4.00 input / $20.00 output per million tokens (official OpenAI API price). At a typical mix of three input tokens to one output token, that is $4.00 per million tokens for Claude Sonnet 5.5 versus $6.00 for Gemini 4 Neon (1.5× as much) and $8.00 for GPT-5.6 Sol (2× as much).
Which scores higher on benchmarks?
Not every model here has a Capabilities Index score, so we compare the average of shared benchmarks (GPQA Diamond, FrontierMath Tiers 1–3, OTIS Mock AIME 2024–2025 and SimpleQA Verified): Gemini 4 Neon 90.7%, GPT-5.6 Sol 88.1% and Claude Sonnet 5.5 82.7%. On individual benchmarks: GPQA Diamond — Gemini 4 Neon 96.3%, Claude Sonnet 5.5 95.6%, GPT-5.6 Sol 93.5%; FrontierMath Tiers 1–3 — Gemini 4 Neon 91.2%, GPT-5.6 Sol 89.1%, Claude Sonnet 5.5 88.8%; OTIS Mock AIME 2024–2025 — Gemini 4 Neon 100%, GPT-5.6 Sol 100%, Claude Sonnet 5.5 100%; SimpleQA Verified — Gemini 4 Neon 75.2%, GPT-5.6 Sol 69.7%, Claude Sonnet 5.5 46.5%.
Which is better for coding?
There are no published SWE-bench Verified results for Gemini 4 Neon, GPT-5.6 Sol and Claude Sonnet 5.5 yet, so there is no like-for-like coding score. On overall capability, Gemini 4 Neon 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-5.6 Sol and Gemini 4 Neon have the largest context windows (1,050,000 and 1,048,576 tokens), against 1,000,000 for Claude Sonnet 5.5. Maximum output per response: Gemini 4 Neon up to 65,536, GPT-5.6 Sol up to 128,000, Claude Sonnet 5.5 up to 128,000 tokens.
Which can read images, PDFs, audio or video?
Gemini 4 Neon accepts text, images, PDFs, audio and video; GPT-5.6 Sol accepts text, images and PDFs; Claude Sonnet 5.5 accepts text, images and PDFs. Gemini 4 Neon handles the widest range of inputs.
Are any of these open source?
No. Gemini 4 Neon, GPT-5.6 Sol and Claude Sonnet 5.5 are proprietary and only available through APIs and apps.
Which is newer?
Gemini 4 Neon is the newest, released Nov 18, 2026. Claude Sonnet 5.5 came out Sep 28, 2026; GPT-5.6 Sol came out Jul 9, 2026. Knowledge cutoff: Gemini 4 Neon Mar 2026, GPT-5.6 Sol Feb 16, 2026, Claude Sonnet 5.5 Jun 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.