GPT-5.4 vs Mistral Medium 3.5
GPT-5.4 leads Mistral Medium 3.5 on coding, 64.4 to 25.5 on the Vibe Coding Index, and Mistral Medium 3.5 costs 2.0x less per output token ($7.50 vs $15.00 per million). Updated October 21, 2026; 72 benchmarks have a result for both.
GPT-5.4 or Mistral Medium 3.5: which to pick
Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 25.5 for Mistral Medium 3.5.
Pick Mistral Medium 3.5 for output tokens 2.0x cheaper ($7.50 vs $15.00 per million) and work measured by τ²-Bench Telecom, where it leads.
- Shared results
- 72
- GPT-5.4 only
- 38
- Mistral Medium 3.5 only
- 9
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
GPT-5.4
GPT-5.4 scores 64.4 against 25.5.
3 benchmarks behind each score
Intelligence
AA Intelligence Index
GPT-5.4
GPT-5.4 leads, 39.0 against 14.2.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Mistral Medium 3.5 has a AA Agentic Index result, 7.0.
Directional only
Reasoning
GPQA Diamond
GPT-5.4
GPT-5.4 leads, 91.7% against 34.9%.
Same board, both measured
Math
AIME
No call
Only GPT-5.4 has a AIME result, 96.7%.
Directional only
Knowledge
MMLU Pro
GPT-5.4
GPT-5.4 leads, 87.5% against 75.3%.
Same board, both measured
Web design
Design Arena: websites
GPT-5.4
GPT-5.4 leads, 1199 against 1142.
Same board, both measured
Long documents
Listed context window
GPT-5.4
GPT-5.4 takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Mistral Medium 3.5
GPT-5.4 $0.010, Mistral Medium 3.5 $0.0053 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-5.4
GPT-5.4 $1.02, Mistral Medium 3.5 $6.02.
Published attempt cost and pass rate
At a glance
- Coding index
- GPT-5.4 64.4 from 3 benchmarks, Mistral Medium 3.5 25.5 from 3 benchmarks.
- Widest category
- Coding, 25.9 points apart on average over 11 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): GPT-5.4 67.4% against Mistral Medium 3.5 2.9%, 64.5 points apart on a 0 to 100 scale.
- Output price
- GPT-5.4 $15.00 against Mistral Medium 3.5 $7.50 per million output tokens, 2.0x apart.
- Context window
- GPT-5.4 takes 1.05M tokens, Mistral Medium 3.5 262K.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | GPT-5.4 | Bars | Mistral Medium 3.5 | Ties |
|---|---|---|---|---|
| Coding | 11 | 0 | ||
| Intelligence | 2 | 0 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 9 | 0 | ||
| Math | 3 | 0 | ||
| Knowledge | 10 | 0 | ||
| Web design | 8 | 0 | ||
| General | 28 | 0 | ||
| All shared boards | 71 | 1 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index110
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| LMArena WebDev | 1464 | 1264 | GPT-5.4 |
| AA Coding Index | 71.1 | 46.9 | GPT-5.4 |
| LiveCodeBench (Vals AI) | 84.1% | Not measured | Only GPT-5.4 measured |
| SWE-bench Verified (Vals AI) | 78.2% | 66.4% | GPT-5.4 |
| Vibe Code Bench (Vals AI) | 67.4% | 2.9% | GPT-5.4 |
| Code migration (Vals AI) | 35.0% | 5.1% | GPT-5.4 |
| ProgramBench (Vals AI) | 0.5% | Not measured | Only GPT-5.4 measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 39.0% | Only Mistral Medium 3.5 measured |
| SkillsBench (Vals AI) | 51.7% | Not measured | Only GPT-5.4 measured |
| SciCode | Not measured | 40.2% | Only Mistral Medium 3.5 measured |
| Terminal-Bench Hard (Artificial Analysis run) | 57.6% | 33.3% | GPT-5.4 |
| Terminal-Bench | 81.8% | Not measured | Only GPT-5.4 measured |
| GSO-Bench | 31.4% | Not measured | Only GPT-5.4 measured |
| WeirdML | 77.7% | Not measured | Only GPT-5.4 measured |
| MirrorCode | 15.6% | Not measured | Only GPT-5.4 measured |
| ALE-Bench | 1607.0 | 764.0 | GPT-5.4 |
| LMArena Text: coding | 1496 | 1460 | GPT-5.4 |
| LMArena Text: software and IT | 1491 | 1458 | GPT-5.4 |
| LMArena WebDev: React | 1454 | 1254 | GPT-5.4 |
| LMArena WebDev: HTML | 1478 | 1308 | GPT-5.4 |
| LMArena WebDev: from an image | 1436 | Not measured | Only GPT-5.4 measured |
| Neither model has a result on the other 18 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| AA Intelligence Index | 39.0 | 14.2 | GPT-5.4 |
| Epoch Capabilities Index | 157.3 | 141.4 | GPT-5.4 |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 7.0 | Only Mistral Medium 3.5 measured |
| Design Arena: agents building web apps | 1059 | Not measured | Only GPT-5.4 measured |
| Design Arena: agents building full-stack apps | 1017 | Not measured | Only GPT-5.4 measured |
| Design Arena: agents building mobile apps | 1087 | Not measured | Only GPT-5.4 measured |
| Design Arena: agents building Android apps | 1030 | Not measured | Only GPT-5.4 measured |
| Design Arena: agents building Godot games | 1135 | Not measured | Only GPT-5.4 measured |
| Finance Agent (Vals AI) | Not measured | 32.1% | Only Mistral Medium 3.5 measured |
| Tax Agent Bench (Vals AI) | Not measured | 4.5% | Only Mistral Medium 3.5 measured |
| Legal Research Bench (Vals AI) | Not measured | 9.1% | Only Mistral Medium 3.5 measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 11.2% | Only Mistral Medium 3.5 measured |
| τ²-Bench Telecom | 87.1% | 94.2% | Mistral Medium 3.5 |
| APEX-Agents | 52.4% | Not measured | Only GPT-5.4 measured |
| METR time horizons | 74.3% | Not measured | Only GPT-5.4 measured |
| DeepResearch Bench | 35.1% | Not measured | Only GPT-5.4 measured |
| PostTrainBench | 19.0% | Not measured | Only GPT-5.4 measured |
| Vending-Bench 2 | $6,144 | Not measured | Only GPT-5.4 measured |
| Neither model has a result on the other 25 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index90
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| GPQA Diamond | 91.7% | 34.9% | GPT-5.4 |
| MMMU Pro (Vals AI) | 87.5% | Not measured | Only GPT-5.4 measured |
| GPQA Diamond (Artificial Analysis run) | 92.0% | 74.8% | GPT-5.4 |
| Humanity's Last Exam (Artificial Analysis run) | 43.7% | 13.8% | GPT-5.4 |
| GPQA Diamond (Epoch AI run) | 93.3% | Not measured | Only GPT-5.4 measured |
| ARC-AGI | 93.7% | Not measured | Only GPT-5.4 measured |
| ARC-AGI-2 | 74.0% | Not measured | Only GPT-5.4 measured |
| Chess puzzles | 48.0% | Not measured | Only GPT-5.4 measured |
| Mystery game puzzles | 37.0% | Not measured | Only GPT-5.4 measured |
| EBR-bench | 25.4% | Not measured | Only GPT-5.4 measured |
| LMArena Text: hard prompts | 1485 | 1436 | GPT-5.4 |
| LMArena Text: expert | 1507 | 1432 | GPT-5.4 |
| LMArena Text: hard prompts in English | 1479 | 1445 | GPT-5.4 |
| LMArena Vision: diagrams | 1327 | 1254 | GPT-5.4 |
| LMArena Vision: homework | 1336 | 1230 | GPT-5.4 |
| Kagi LLM Benchmark | 63.8% | 41.4% | GPT-5.4 |
| Neither model has a result on the other 6 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| AIME | 96.7% | Not measured | Only GPT-5.4 measured |
| OTIS Mock AIME | 97.8% | Not measured | Only GPT-5.4 measured |
| FrontierMath, tiers 1 to 3 | 78.6% | Not measured | Only GPT-5.4 measured |
| FrontierMath, tier 4 | 49.0% | Not measured | Only GPT-5.4 measured |
| ProofBench | 56.0% | 9.0% | GPT-5.4 |
| LMArena Text: math | 1488 | 1431 | GPT-5.4 |
| LMArena Text: mathematical work | 1497 | 1440 | GPT-5.4 |
| Neither model has a result on the other 6 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index100
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| MMLU Pro | 87.5% | 75.3% | GPT-5.4 |
| LegalBench (Vals AI) | 86.0% | Not measured | Only GPT-5.4 measured |
| CaseLaw (Vals AI) | 63.8% | 44.2% | GPT-5.4 |
| CorpFin (Vals AI) | 65.3% | 58.8% | GPT-5.4 |
| TaxEval (Vals AI) | 74.0% | 68.0% | GPT-5.4 |
| MortgageTax (Vals AI) | 68.3% | 28.9% | GPT-5.4 |
| MedQA (Vals AI) | 96.1% | Not measured | Only GPT-5.4 measured |
| MedCode (Vals AI) | 41.3% | 33.8% | GPT-5.4 |
| SimpleQA Verified | 45.1% | Not measured | Only GPT-5.4 measured |
| LMArena Text: science | 1479 | 1428 | GPT-5.4 |
| LMArena Text: medicine and healthcare | 1458 | 1409 | GPT-5.4 |
| LMArena Text: legal and government | 1486 | 1437 | GPT-5.4 |
| LMArena Text: business and finance | 1474 | 1426 | GPT-5.4 |
| Neither model has a result on the other 5 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index80
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| Design Arena: websites | 1199 | 1142 | GPT-5.4 |
| Design Arena: UI components | 1215 | 1079 | GPT-5.4 |
| Design Arena: game dev | 1195 | 1079 | GPT-5.4 |
| Design Arena: data viz | 1196 | 1120 | GPT-5.4 |
| Design Arena: 3D | 1066 | 1061 | GPT-5.4 |
| Design Arena: all code categories | 1180 | 1121 | GPT-5.4 |
| Design Arena: SVG | 1142 | 950 | GPT-5.4 |
| Design Arena: ASCII art | 1188 | 1063 | GPT-5.4 |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index280
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| SAGE (Vals AI) | 43.3% | 37.6% | GPT-5.4 |
| MedScribe (Vals AI) | 77.5% | 67.7% | GPT-5.4 |
| Vals Multimodal Index | Not measured | 34.8% | Only Mistral Medium 3.5 measured |
| IFBench | 73.9% | 68.8% | GPT-5.4 |
| Long context reasoning (AA-LCR) | 82.0% | 69.3% | GPT-5.4 |
| DTBench | 94.4% | 75.5% | GPT-5.4 |
| LMCA | 52.0% | 26.1% | GPT-5.4 |
| CL-bench | 27.9% | Not measured | Only GPT-5.4 measured |
| Surface Evolver Bench | Not measured | 26.9% | Only Mistral Medium 3.5 measured |
| Furniture assembly | 37.5% | Not measured | Only GPT-5.4 measured |
| CL-bench Life | 21.7% | Not measured | Only GPT-5.4 measured |
| LMArena Text | 1469 | 1421 | GPT-5.4 |
| LMArena Text: instruction following | 1469 | 1415 | GPT-5.4 |
| LMArena Text: creative writing | 1439 | 1374 | GPT-5.4 |
| LMArena Vision | 1303 | 1222 | GPT-5.4 |
| LMArena Text: English | 1467 | 1433 | GPT-5.4 |
| LMArena Text: non-English | 1465 | 1404 | GPT-5.4 |
| LMArena Text: Chinese | 1520 | 1444 | GPT-5.4 |
| LMArena Text: German | 1472 | 1453 | GPT-5.4 |
| LMArena Text: Japanese | 1485 | Not measured | Only GPT-5.4 measured |
| LMArena Text: Korean | 1449 | 1384 | GPT-5.4 |
| LMArena Text: French | 1493 | 1448 | GPT-5.4 |
| LMArena Text: Spanish | 1452 | 1411 | GPT-5.4 |
| LMArena Text: Russian | 1480 | 1395 | GPT-5.4 |
| LMArena Text: Polish | 1472 | 1430 | GPT-5.4 |
| LMArena Text: multi-turn | 1482 | 1423 | GPT-5.4 |
| LMArena Text: longer queries | 1473 | 1415 | GPT-5.4 |
| LMArena Text: writing and language | 1458 | 1384 | GPT-5.4 |
| LMArena Text: entertainment and media | 1438 | 1382 | GPT-5.4 |
| LMArena Vision: English | 1301 | 1240 | GPT-5.4 |
| LMArena Vision: Chinese | 1354 | 1244 | GPT-5.4 |
| LMArena Vision: reading text | 1312 | 1233 | GPT-5.4 |
| LMArena Vision: humor | 1284 | 1216 | GPT-5.4 |
| LMArena Vision: recognizing things | 1251 | Not measured | Only GPT-5.4 measured |
| LMArena Vision: creative writing | 1283 | 1198 | GPT-5.4 |
| LMArena Document | 1471 | Not measured | Only GPT-5.4 measured |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | GPT-5.4 | Mistral Medium 3.5 | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Mistral Medium 3.5 costs less
Repository review
50K input + 3K output tokens
Mistral Medium 3.5 costs less
Agent loop
200K input + 10K output tokens
Mistral Medium 3.5 costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | OpenAI GPT-5.4 | Mistral Medium 3.5 |
|---|---|---|
| Maker | OpenAI | Mistral AI |
| Weights | No public weights listed | No public weights listed |
| Released | March 5, 2026 | April 30, 2026 |
| Context window | 1.05M | 262K |
| Input $/Mtok | $2.50 | $1.50 |
| Output $/Mtok | $15.00 | $7.50 |
| Coding index | 64.4 | 25.5 |
| Independent measurements behind it | 3 | 3 |
| Cost per solved task | $1.02 | $6.02 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is GPT-5.4 better than Mistral Medium 3.5 for coding?
Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 25.5 for Mistral Medium 3.5. Pick Mistral Medium 3.5 for output tokens 2.0x cheaper ($7.50 vs $15.00 per million) and work measured by τ²-Bench Telecom, where it leads.
Which is cheaper, GPT-5.4 or Mistral Medium 3.5?
GPT-5.4 costs $2.50 per million input tokens and $15.00 per million output tokens; Mistral Medium 3.5 costs $1.50 and $7.50. Per solved SWE-bench Verified task, GPT-5.4 costs $1.02 and Mistral Medium 3.5 $6.02.
Which has the bigger context window, GPT-5.4 or Mistral Medium 3.5?
GPT-5.4, with 1.05M tokens against 262K.
How many benchmarks is the GPT-5.4 vs Mistral Medium 3.5 comparison based on?
72 benchmarks have a result for both models as of October 21, 2026. GPT-5.4's coding index rests on 3 independent benchmarks and Mistral Medium 3.5's on 3. Where only one model has a result, the page says so instead of naming a winner.