Across 15 comparable benchmarks, Gemini 4 Neon has better reported results on 14 and worse on 1. 2 further shared results are listed below, each with the reason it is not counted.
Moderate evidence15 of 17 shared results countedHow we compare
Price basis: Peak-hour rates vs standard rates. The price section below shows the applicable periods.
| Fact | DeepSeek-V4.1-Flash-1017 | Gemini 4 Neon |
|---|---|---|
| Context window | 1M tokens | 1M tokens |
| Input / 1M | $0.300 | $3.00 |
| Output / 1M | $1.20 | $15.00 |
| Blended (4:1) | $0.268 | $3.35 |
| Status | Available | Available |
| Released | October 17, 2026 | November 18, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates β set your own mix.
Mean of 7 shared benchmarks grouped by capability on a 0β100 scale. How averages are built
Gemini 4 Neon ahead on 5 of 5 counted
Gemini 4 Neon ahead on 1 of 1 counted
Split 1β1 of 2 counted
Gemini 4 Neon ahead on 3 of 3 counted
No shared comparable results for Abstract reasoning, Mathematics, Science & health, Research & search, Long context, Cybersecurity.
Published scores, with matching modes where available. Gaps show DeepSeek-V4.1-Flash-1017 relative to Gemini 4 Neon. Reading the results
| Benchmark | DeepSeek-V4.1-Flash-1017 | Gemini 4 Neon | Gap |
|---|---|---|---|
| FrontierCode 2.0Repository Engineering Β· Weighted rubric score | 19.3Unspecified Reasoning Β· tools | 46.2β Thinking High Β· tools + web | 0.42Γ#23 vs #3 of 32 |
| SWE-sweepRepository Engineering Β· Bugs resolved (%) | 1.7Max Β· tools | 12.8β Extra-High Β· tools | β11.1 pts#13 vs #2 of 20 |
| Terminal-Bench 4.0Agentic Development Β· Resolution rate (%) | 42.3Unspecified Reasoning Β· tools | 63.6β Extra-High Β· tools | β21.3 pts#25 vs #7 of 90 |
| Terminal-Bench 4.1Agentic Development Β· Resolution rate (%) | 38.5Max Β· tools | 65.2β Extra-High Β· tools | β26.7 pts#21 vs #3 of 48 |
| Vibe Code Bench v1.1Code Generation & Editing Β· Pass rate (%) | 87.7Max Β· tools | 93.7β Extra-High Β· tools | β6.0 pts#15 vs #2 of 119 |
| Agents' Last Exam v1.1Tool Orchestration Β· Pass rateNot counted Β· Different or unrecorded test environments β excluded from the tally | 36.1Thinking High Β· toolsΒ· pass-rate | 75.8β Extra-High Β· toolsΒ· score | β#14 vs #1 of 21 |
| AutomationBench-AATool Orchestration Β· Guardrail-safe objective completion rate | 66.6Max Β· tools | 79.3β Extra-High Β· tools | β12.7 pts#12 vs #1 of 15 |
| ChartographyDocuments & Charts Β· Score | 83.0Unspecified Reasoning Β· tools | 73.5β Extra-High Β· no tools | 1.13Γ#9 vs #15 of 24 |
| GDP.pdfDocuments & Charts Β· Task pass rate (%) | 17.2Max Β· no tools | 33.4β Thinking High Β· no tools | β16.2 pts#37 vs #3 of 93 |
| AA-BriefcaseOffice & Business Β· AA-Briefcase Elo | 1570Max Β· tools | 1700β Extra-High Β· tools | 0.92Γ#14 vs #5 of 17 |
| AutomationBenchOffice & Business Β· Pass RateNot counted Β· Different or unrecorded test environments β excluded from the tally | 60.5Unspecified Reasoning Β· toolsΒ· legacy-default | 56.3β Extra-High Β· toolsΒ· zapier-private-heldout | β#2 vs #3 of 64 |
| Legal Research BenchLegal Β· Score | 45.2Max Β· tools | 56.3β Extra-High Β· tools + internet | 0.80Γ#21 vs #1 of 91 |
| Tax Agent BenchFinance Β· Accuracy (%, rubric Γ citation quality) | 44.4Max Β· tools | 54.9β Extra-High Β· tools + internet | β10.5 pts#26 vs #22 of 74 |
| Reported separately from the capability domains; not included in the averages. Read each metricβs direction β lower can be better. | |||
| AA-OmniscienceHonesty & Factuality Β· AA-Omniscience Index | 4.7Max Β· no tools | 54.8β Extra-High Β· no tools | 0.09Γ#16 vs #1 of 16 |
| HLE-DiamondKnowledge Exams Β· Accuracy (%) | 21.6Max Β· no tools | 71.3β Extra-High Β· no tools | β49.7 pts#11 vs #3 of 11 |
| AA Intelligence Index v5.0.1Capability Indices Β· Index (0-100) | 32.6Max Β· task tools | 61.4β Extra-High Β· tools | 0.53Γ#25 vs #3 of 61 |
| Vals Index v2.2Capability Indices Β· GDP-weighted accuracy (%) | 61.3Max Β· task tools | 70.9β Extra-High Β· tools | β9.6 pts#15 vs #2 of 34 |
Repository Engineering Β· Weighted rubric score
Repository Engineering Β· Bugs resolved (%)
Agentic Development Β· Resolution rate (%)
Agentic Development Β· Resolution rate (%)
Code Generation & Editing Β· Pass rate (%)
Tool Orchestration Β· Pass rate
Not counted Β· Different or unrecorded test environments β excluded from the tally
Tool Orchestration Β· Guardrail-safe objective completion rate
Documents & Charts Β· Score
Documents & Charts Β· Task pass rate (%)
Office & Business Β· AA-Briefcase Elo
Office & Business Β· Pass Rate
Not counted Β· Different or unrecorded test environments β excluded from the tally
Legal Β· Score
Finance Β· Accuracy (%, rubric Γ citation quality)
Honesty & Factuality Β· AA-Omniscience Index
Knowledge Exams Β· Accuracy (%)
Capability Indices Β· Index (0-100)
Capability Indices Β· GDP-weighted accuracy (%)
A dash means no result is available.
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix Β· 5% steps
Blended price = effective input rate at 90% cache hits Γ 80% + output rate Γ 20%
DeepSeek-V4.1-Flash-1017 costs92%lessthan Gemini 4 Neon per million tokens β 12.5Γ cheaper.Using peak-hour rates vs standard rates at this token mix.
DeepSeek-V4.1-Flash-1017: Peak-hour pricing. DeepSeek official USD pricing (peak: 01:00β04:00 and 06:00β10:00 UTC MonβFri exc. CN holidays) DeepSeek official USD pricing (peak output)
Gemini 4 Neon: Standard pricing. Unified 1M context pricing; thinking tokens included in output; paid tier only (no free tier).
| Price / 1M tokens | DeepSeek-V4.1-Flash-1017 | Gemini 4 Neon | Ratio |
|---|---|---|---|
| Input | $0.30 | $3.00 | 0.10Γ |
| Output | $1.20 | $15.00 | 0.08Γ |
| Cached input | $0.006 | $0.15 | 0.04Γ |
| Cache write | β | β | β |
| Blended | $0.268 | $3.35 | 0.08Γ |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
DeepSeek-V4.1-Flash-1017 relative to Gemini 4 Neon
Summary: 1 of 5 shared benchmarks have matching recorded settings and sources. The average is a geometric mean of 1 positive-cost ratios, one equal weight per benchmark. This benchmark set can include related tests and indexes; it is not a workload cost estimate.
DeepSeek-V4.1-Flash-1017 8.2Γ cheaper
Settings or sources differ
Settings or sources differ
Settings or sources differ
Settings or sources differ
DeepSeek-V4.1-Flash-1017 8.2Γ cheaper
Settings or sources differ
Settings or sources differ
Settings or sources differ
Settings or sources differ
Effort, protocol, aggregation and source match; other run conditions may differ.
DeepSeek-V4.1-Flash-1017
Gemini 4 Neon
Settings or sources differ or are not recorded; interpret the comparison with care.
DeepSeek-V4.1-Flash-1017
Gemini 4 Neon
Settings or sources differ or are not recorded; interpret the comparison with care.
DeepSeek-V4.1-Flash-1017
Gemini 4 Neon
Settings or sources differ or are not recorded; interpret the comparison with care.
DeepSeek-V4.1-Flash-1017
Gemini 4 Neon
Settings or sources differ or are not recorded; interpret the comparison with care.
DeepSeek-V4.1-Flash-1017
Gemini 4 Neon
| Benchmark | DeepSeek-V4.1-Flash-1017 | Gemini 4 Neon |
|---|---|---|
| AA Coding Agent Index v2.0 | Not reported | |
| AA Intelligence Index v4.3 | Not reported | |
| AA-Briefcase | Not reported | |
| AA-LCR | Not reported | |
| AA-Omniscience | Not reported | |
| Agent Arena | $0.1402 / task Median Β· Net improvement index (percentage points vs. average model): 2 Max (With Tools + Internet) | Not reported |
| ARC-AGI-1 | 1 more settings | Not reported |
| ARC-AGI-2 | 1 more settings | Not reported |
| AutomationBench-AA | $0.1739 / task Average Β· Guardrail-safe objective completion rate: 66.6 Max (With Tools) | Not reported |
| Code Migration | Not reported | |
| CritPt | Not reported | |
| CursorBench 4.0 | $0.98 / task Average Β· Task score: 47.7 Max (With Tools) Β· cursorbench-4-0-cursor-agent $0.77 / task Average Β· Task score: 45.6 Thinking High (With Tools) Β· cursorbench-4-0-cursor-agent $0.60 / task Average Β· Task score: 44.9 Thinking Low (With Tools) Β· cursorbench-4-0-cursor-agent 1 more settings$0.80 / task Average Β· Task score: 46.3 Thinking Medium (With Tools) Β· cursorbench-4-0-cursor-agent | Not reported |
| EMB (Excel Modeling) | Not reported | |
| Finance Agent v2 | Not reported | |
| FrontierCS 2 | Not reported | |
| GDP.pdf | Not reported | |
| GDPval-AA v2 | Not reported | |
| GDPx | $0.94 / task Average Β· GDPx β Score: 5.4 Highest (source-reported) (With Tools) Β· overall/score | Not reported |
| Harvey's Legal Agent Benchmark | Not reported | |
| HLE | Not reported | |
| LiveBench 2.0 | $0.1305 / task Average Β· Overall score (equal-weight mean of 7 suite scores, 0-100): 79.2 Max (task-specific tools) | Not reported |
| SciCode | Not reported | |
| SWE-sweep | Not reported | |
| Terminal-Bench 4.0 | $0.54 / task Average Β· Resolution rate (%): 31.8 Unspecified Reasoning (With Tools) Β· vals-mini-swe | Not reported |
| Terminal-Bench 4.1 | Not reported | |
| Vals Index v2.1 | Not reported | |
| WeirdML v3 | $0.879 / task Average Β· Official score (0-1; higher is better): 0.092 Thinking High (With Tools) | Not reported |
| Attribute | DeepSeek-V4.1-Flash-1017 | Gemini 4 Neon |
|---|---|---|
| Availability | Available | Available |
| Provider | DeepSeek-AI | Google Deep Mind |
| Model type | Multimodal model | Multimodal model |
| First release | October 17, 2026 | November 18, 2026 |
| Context window | 1M tokens | 1M tokens |
| Max output | 384K tokens | 1M tokens |
| Total parameters | 552B | β |
| Active parameters | 16B | β |
| Architecture | Mixture of Experts (MoE) | Undisclosed |
| Documented capabilities | Reasoning, Multilingual, Coding, Function calling, Tool use, Structured output, Long context | Reasoning, Coding, Tool use, Computer use, Long context, Function calling, Structured output, Web search |
| Knowledge cutoff | β | 2026-03 |
| Input modalities | Text, Image | Text, Image, Audio, Video |
| Output modalities | Text | Text |
| Weights | Open weights | Closed / API only |
| Licence | MIT License | β |
| Commercial use | Commercial use allowed | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used DeepSeek-V4.1-Flash-1017 or Gemini 4 Neon? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.