Across 15 comparable benchmarks, Gemini 4 Argon has better reported results on 15 and worse on 0.
Strong evidence15 of 15 shared results countedHow we compare
Price basis: Current effective rates. The price section below shows the applicable periods.
| Fact | Gemini 4 Argon | Mistral Medium 3.5 |
|---|---|---|
| Context window | 1M tokens | 256K tokens |
| Input / 1M | $2.00 | $1.50 |
| Output / 1M | $10.00 | $7.50 |
| Blended (4:1) | $1.12 | $2.70 |
| Status | Available | Available |
| Released | October 13, 2026 | May 1, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates — set your own mix.
Mean of 11 shared benchmarks grouped by capability on a 0–100 scale. How averages are built
Gemini 4 Argon ahead on 3 of 3 counted
Gemini 4 Argon ahead on 1 of 1 counted
Gemini 4 Argon ahead on 3 of 3 counted
Gemini 4 Argon ahead on 7 of 7 counted
No shared comparable results for Abstract reasoning, Computer & tool use, Research & search, Long context, Multimodal, Cybersecurity.
Published scores, with matching modes where available. Gaps show Gemini 4 Argon relative to Mistral Medium 3.5. Reading the results
| Benchmark | Gemini 4 Argon | Mistral Medium 3.5 | Gap |
|---|---|---|---|
| Code MigrationMaintenance & Optimization · Hidden-test pass rate (%) | 68.2Thinking High · tools | 5.1Thinking High · tools | +63.0 pts#3 vs #72 of 79 |
| FrontierCode 1.1Code Generation & Editing · Weighted rubric score (%) | 47.5Thinking Medium · tools· frontiercode-1-1-main | 8.0≠Normal · tools· frontiercode-1-1-main | +39.5 pts#12 vs #51 of 51 |
| Vibe Code Bench v1.1Code Generation & Editing · Pass rate (%) | 91.9Thinking High · tools | 2.9Thinking High · tools | +89.0 pts#4 vs #115 of 121 |
| ProofBench v1.1 (Lean 4)Mathematics · Accuracy (%) | 99.0Thinking High · tools | 9.0Thinking High · tools | +90.0 pts#6 vs #50 of 54 |
| SciCodeScientific Computing · Score | 61.8Thinking High · no tools | 40.2≠Thinking · no tools | 1.54×#5 vs #75 of 98 |
| MedCodeClinical Workflows · Accuracy (%) | 58.8Thinking High · tools | 33.8Thinking High · tools | +25.0 pts#3 vs #78 of 97 |
| MedScribeClinical Workflows · Rubric pass rate (%) | 87.4Thinking High · tools | 67.7Thinking High · tools | +19.7 pts#17 vs #90 of 99 |
| AA-AnalystAgentData Analysis · pass^5 (%) | 48.8Thinking High · tools + web | 12.5≠Normal · tools + web | +36.3 pts#13 vs #34 of 39 |
| EMB (Excel Modeling)Finance · Accuracy (%) | 75.2Thinking High · tools | 11.2Thinking High · tools | +64.0 pts#6 vs #76 of 78 |
| Finance Agent v2Finance · Score | 65.4Max · tools | 32.1≠Thinking High · tools | 2.04×#1 vs #67 of 75 |
| Harvey's Legal Agent BenchmarkLegal · Accuracy (%) | 19.6Thinking High · tools | 0.42Thinking High · tools | +19.2 pts#4 vs #43 of 52 |
| Legal Research BenchLegal · Score | 54.8Thinking High · tools | 9.1Thinking High · tools | 6.00×#6 vs #83 of 91 |
| SAGEService Workflows · Accuracy (%) | 53.6Thinking High · tools | 37.6Thinking High · tools | +16.0 pts#5 vs #62 of 82 |
| Tax Agent BenchFinance · Accuracy (%, rubric × citation quality) | 44.6Thinking High · tools | 4.5Thinking High · tools | +40.1 pts#25 vs #73 of 74 |
| AA Intelligence Index v5.0.1Capability Indices · Index (0-100) | 50.4Thinking High · tools | 4.6Thinking High · tools | 10.9×#11 vs #59 of 62 |
Maintenance & Optimization · Hidden-test pass rate (%)
Code Generation & Editing · Weighted rubric score (%)
Code Generation & Editing · Pass rate (%)
Mathematics · Accuracy (%)
Data Analysis · pass^5 (%)
Finance · Accuracy (%)
Finance · Score
Legal · Accuracy (%)
Legal · Score
Service Workflows · Accuracy (%)
Finance · Accuracy (%, rubric × citation quality)
Capability Indices · Index (0-100)
A dash means no result is available.
Both models publish peak and off-peak prices. Each period is compared separately using the same token mix; hours follow each provider’s schedule.
Drag the handle above the bar for cache hits, the one below it for the input : output mix · 5% steps
Blended price = effective input rate at 90% cache hits × 80% + output rate × 20%
Gemini 4 Argon costs17%lessthan Mistral Medium 3.5 per million tokens.Using current effective rates at this token mix.
Gemini 4 Argon: Listed text rates; pricing tier not confirmed.
Mistral Medium 3.5: Standard pricing. Mistral official pricing verified 2026-11-13; Deprecated 2026-10-31, scheduled retirement 2027-04-30 per docs.mistral.ai/models
| Price / 1M tokens | Gemini 4 Argon | Mistral Medium 3.5 | Ratio |
|---|---|---|---|
| Input | $2.00 | $1.50 | 1.33× |
| Output | $10.00 | $7.50 | 1.33× |
| Cached input | $0.10 | — | — |
| Cache write | — | — | — |
| Blended | $2.23 | $2.70 | 0.83× |
Gemini 4 Argon costs65%morethan Mistral Medium 3.5 per million tokens.Using standard list rates after promotions end at this token mix.
Gemini 4 Argon: Standard pricing. Google Gemini API official pricing (ai.google.dev/gemini-api/docs/pricing) verified 2026-10-21; paid tier; standard tier; scheduled price starting 2026-12-01
Mistral Medium 3.5: Standard pricing. Mistral official pricing verified 2026-11-13; Deprecated 2026-10-31, scheduled retirement 2027-04-30 per docs.mistral.ai/models
| Price / 1M tokens | Gemini 4 Argon | Mistral Medium 3.5 | Ratio |
|---|---|---|---|
| Input | $4.00 | $1.50 | 2.67× |
| Output | $20.00 | $7.50 | 2.67× |
| Cached input | $0.20 | — | — |
| Cache write | — | — | — |
| Blended | $4.46 | $2.70 | 1.65× |
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.
Gemini 4 Argon relative to Mistral Medium 3.5
Summary: 2 of 2 shared benchmarks have matching recorded settings and sources. The average is a geometric mean of 2 positive-cost ratios, one equal weight per benchmark. This benchmark set can include related tests and indexes; it is not a workload cost estimate.
Gemini 4 Argon 9.6× more expensive
Gemini 4 Argon 2.4× more expensive
Gemini 4 Argon 9.6× more expensive
Gemini 4 Argon 2.4× more expensive
Effort, protocol, aggregation and source match; other run conditions may differ.
Gemini 4 Argon
Mistral Medium 3.5
Effort, protocol, aggregation and source match; other run conditions may differ.
Gemini 4 Argon
Mistral Medium 3.5
| Benchmark | Gemini 4 Argon | Mistral Medium 3.5 |
|---|---|---|
| AA Coding Agent Index v1.5 | Not reported | |
| AA Coding Agent Index v2.0 | Not reported | |
| AA Intelligence Index v4.3 | Not reported | |
| AA-Omniscience | Not reported | |
| Agent Arena | $0.694 / task Median · Net improvement index (percentage points vs. average model): 10.7 Thinking High (With Tools + Internet) | Not reported |
| Agent Arena Index | $0.6569 / task Median · Agent Arena Index rating (index points): 1229 Thinking High (With Tools + Internet) | Not reported |
| ARC-AGI-1 | Not reported | |
| ARC-AGI-2 | Not reported | |
| AutomationBench | $1.70 / task Average · Pass Rate: 51.3 Thinking High (With Tools) · zapier-private-heldout $1.54 / task Average · Pass Rate: 50.1 Thinking Medium (With Tools) · zapier-private-heldout | Not reported |
| AutomationBench-AA | $0.6087 / task Average · Guardrail-safe objective completion rate: 77.5 Thinking High (With Tools) | Not reported |
| CritPt | Not reported | |
| CursorBench 4.0 | $4.57 / task Average · Task score: 52.9 Thinking High (With Tools) · cursorbench-4-0-cursor-agent $4.09 / task Average · Task score: 50.6 Thinking Low (With Tools) · cursorbench-4-0-cursor-agent $4.39 / task Average · Task score: 52.8 Thinking Medium (With Tools) · cursorbench-4-0-cursor-agent | Not reported |
| CWE-bench v1 | Not reported | |
| FrontierCode 2.0 | $6.514 / task Average · Weighted rubric score: 26.3 Thinking High (With Tools + Internet) $5.688 / task Average · Weighted rubric score: 23.4 Thinking Medium (With Tools + Internet) | Not reported |
| FrontierCS 2 | Not reported | |
| FrontierSWE v2 | Not reported | |
| GDPx | Not reported | |
| HLE | Not reported | |
| LiveBench 2.0 | $1.5159 / task Average · Overall score (equal-weight mean of 7 suite scores, 0-100): 84.1 Thinking High (With Tools) | Not reported |
| MedCode | Not reported | |
| MedScribe | Not reported | |
| MysteryMechanism | Not reported | |
| ProofBench v1.1 (Lean 4) | Not reported | |
| Public Benefits Bench v1.1 | Not reported | |
| SAGE | Not reported | |
| SciCode | Not reported | |
| Terminal-Bench 4.0 | $16.8269 / task Average · Resolution rate (%): 55.6 Thinking High (With Tools) · vals-effort-study $13.4495 / task Average · Resolution rate (%): 52.5 Thinking Low (With Tools) · vals-effort-study 2 more settings$13.3909 / task Average · Resolution rate (%): 56.6 Thinking Medium (With Tools) · vals-effort-study $17.64 / task Average · Resolution rate (%): 57.6 Unspecified Reasoning (With Tools) · vals-mini-swe | Not reported |
| Terminal-Bench 4.1 | Not reported | |
| Terminal-Bench-Science 0.1 | Not reported | |
| Vals Index v2.1 | Not reported | |
| Vals Index v2.2 | Not reported | |
| Vibe Code Bench 1-100 | Not reported | |
| Vibe Code Bench v1.1 | Not reported | |
| WeirdML v3 | $16.2075 / task Average · Official score (0-1; higher is better): 0.2 Thinking High (With Tools) | Not reported |
| Attribute | Gemini 4 Argon | Mistral Medium 3.5 |
|---|---|---|
| Availability | Available | Available |
| Provider | Google Deep Mind | MistralAI |
| Model type | Multimodal model | Chat model |
| First release | October 13, 2026 | May 1, 2026 |
| Context window | 1M tokens | 256K tokens |
| Max output | 1M tokens | — |
| Architecture | Undisclosed | Undisclosed |
| Documented capabilities | Reasoning, Coding, Tool use, Computer use, Long context | Coding, Tool use, Function calling, Long context |
| Input modalities | Text, Image, Audio, Video | Text |
| Output modalities | Text | Text |
| Weights | Closed / API only | Closed / API only |
| Commercial use | Proprietary | 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 Gemini 4 Argon or Mistral Medium 3.5? 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.