Devstral 2 2512 vs GPT-5.6 Terra
GPT-5.6 Terra leads Devstral 2 2512 on the Artificial Analysis Intelligence Index, 42.1 to 8.6, and Devstral 2 2512 costs 6.0x less per output token ($2.00 vs $12.00 per million). Updated October 22, 2026; 5 benchmarks have a result for both.
vs boards won of 5 shared
Devstral 2 2512 or GPT-5.6 Terra: which to pick
Only GPT-5.6 Terra has a coding score, 63.9 on the Vibe Coding Index; Devstral 2 2512 has none, so the two cannot be ranked on coding.
Pick Devstral 2 2512 for output tokens 6.0x cheaper ($2.00 vs $12.00 per million).
Model ADevstral 2 2512Mistral AINot benchmarked1 independent benchmark$1.19 per solved taskModel BGPT-5.6 TerraOpenAI63.9/1006 independent benchmarks$0.42 per solved task
- Shared results
- 5
- Devstral 2 2512 only
- 1
- GPT-5.6 Terra only
- 122
- Like-for-like categories
- 3 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
No call
Only GPT-5.6 Terra has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
GPT-5.6 Terra
GPT-5.6 Terra leads, 42.1 against 8.6.
Same board, both measured
Agentic work
AA Agentic Index
GPT-5.6 Terra
GPT-5.6 Terra leads, 43.2 against 2.3.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only GPT-5.6 Terra has a GPQA Diamond result, 90.9%.
Directional only
Knowledge
MMLU Pro
No call
Only GPT-5.6 Terra has a MMLU Pro result, 86.7%.
Directional only
Web design
Design Arena: websites
No call
Only GPT-5.6 Terra has a Design Arena: websites result, 1268.
Directional only
Long documents
Listed context window
GPT-5.6 Terra
GPT-5.6 Terra takes 1.05M tokens against 262K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Devstral 2 2512
Devstral 2 2512 $0.0014, GPT-5.6 Terra $0.0080 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
GPT-5.6 Terra
Devstral 2 2512 $1.19, GPT-5.6 Terra $0.42.
Published attempt cost and pass rate
At a glance
- Coding index
- Devstral 2 2512 not benchmarked, GPT-5.6 Terra 63.9 from 6 benchmarks.
- Widest category
- Agentic work, 40.9 points apart on average over 1 shared benchmark.
- Widest single gap
- AA Coding Index: Devstral 2 2512 31.3 against GPT-5.6 Terra 76.7, 45.4 points apart on a 0 to 100 scale.
- Output price
- Devstral 2 2512 $2.00 against GPT-5.6 Terra $12.00 per million output tokens, 6.0x apart.
- Context window
- Devstral 2 2512 takes 262K tokens, GPT-5.6 Terra 1.05M.
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 | Devstral 2 2512 | Bars | GPT-5.6 Terra | Ties |
|---|---|---|---|---|
| Coding | 0 | 3 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 0 | 1 | ||
| All shared boards | 0 | 5 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index03
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| SWE-bench Verified | 53.8% | Not measured | Only Devstral 2 2512 measured |
| LiveBench coding | Not measured | 86.2% | Only GPT-5.6 Terra measured |
| LMArena WebDev | Not measured | 1522 | Only GPT-5.6 Terra measured |
| AA Coding Index | 31.3 | 76.7 | GPT-5.6 Terra |
| LiveCodeBench (Vals AI) | 51.8% | 85.9% | GPT-5.6 Terra |
| SWE-bench Verified (Vals AI) | 62.8% | 95.4% | GPT-5.6 Terra |
| Terminal-Bench 4 (Vals AI) | Not measured | 22.7% | Only GPT-5.6 Terra measured |
| Vibe Code Bench (Vals AI) | Not measured | 74.6% | Only GPT-5.6 Terra measured |
| Code migration (Vals AI) | Not measured | 47.8% | Only GPT-5.6 Terra measured |
| ProgramBench (Vals AI) | Not measured | 0.5% | Only GPT-5.6 Terra measured |
| IOI (Vals AI) | Not measured | 87.6% | Only GPT-5.6 Terra measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 77.5% | Only GPT-5.6 Terra measured |
| Vibe Code Bench 1-100 (Vals AI) | Not measured | 14.8% | Only GPT-5.6 Terra measured |
| SkillsBench (Vals AI) | Not measured | 58.9% | Only GPT-5.6 Terra measured |
| CyberBench (Vals AI) | Not measured | 72.1% | Only GPT-5.6 Terra measured |
| SciCode | Not measured | 55.0% | Only GPT-5.6 Terra measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 62.9% | Only GPT-5.6 Terra measured |
| FrontierCode | Not measured | 41.3% | Only GPT-5.6 Terra measured |
| WeirdML | Not measured | 78.3% | Only GPT-5.6 Terra measured |
| ALE-Bench | Not measured | 1951.4 | Only GPT-5.6 Terra measured |
| LMArena Text: coding | Not measured | 1485 | Only GPT-5.6 Terra measured |
| LMArena Text: software and IT | Not measured | 1482 | Only GPT-5.6 Terra measured |
| LMArena WebDev: React | Not measured | 1513 | Only GPT-5.6 Terra measured |
| LMArena WebDev: HTML | Not measured | 1526 | Only GPT-5.6 Terra measured |
| LMArena WebDev: from an image | Not measured | 1541 | Only GPT-5.6 Terra measured |
| SciCode | Not measured | 10.8% | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 13 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| AA Intelligence Index | 8.6 | 42.1 | GPT-5.6 Terra |
| Epoch Capabilities Index | Not measured | 160.1 | Only GPT-5.6 Terra measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| AA Agentic Index | 2.3 | 43.2 | GPT-5.6 Terra |
| Design Arena: agents building web apps | Not measured | 1120 | Only GPT-5.6 Terra measured |
| Design Arena: agents building full-stack apps | Not measured | 1061 | Only GPT-5.6 Terra measured |
| Design Arena: agents building mobile apps | Not measured | 1111 | Only GPT-5.6 Terra measured |
| Design Arena: agents building Android apps | Not measured | 1062 | Only GPT-5.6 Terra measured |
| Design Arena: agents building games | Not measured | 1083 | Only GPT-5.6 Terra measured |
| Design Arena: HTML slides | Not measured | 1058 | Only GPT-5.6 Terra measured |
| Design Arena: PowerPoint slides | Not measured | 1105 | Only GPT-5.6 Terra measured |
| LiveBench research | Not measured | 60.4% | Only GPT-5.6 Terra measured |
| LiveBench browser | Not measured | 84.0% | Only GPT-5.6 Terra measured |
| LiveBench long agentic | Not measured | 48.4% | Only GPT-5.6 Terra measured |
| SREBench (Vals AI) | Not measured | 9.5% | Only GPT-5.6 Terra measured |
| Finance Agent (Vals AI) | Not measured | 54.4% | Only GPT-5.6 Terra measured |
| Tax Agent Bench (Vals AI) | Not measured | 30.0% | Only GPT-5.6 Terra measured |
| Legal Research Bench (Vals AI) | Not measured | 41.4% | Only GPT-5.6 Terra measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 66.2% | Only GPT-5.6 Terra measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 10.0% | Only GPT-5.6 Terra measured |
| τ²-Bench Telecom | Not measured | 86.3% | Only GPT-5.6 Terra measured |
| APEX-Agents | Not measured | 58.2% | Only GPT-5.6 Terra measured |
| BALROG | Not measured | 53.2% | Only GPT-5.6 Terra measured |
| Vending-Bench 2 | Not measured | $7,343 | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 20 boards here | |||
ReasoningDirectional onlyPublic boards, shown as published, kept out of the index17 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 90.9% | Only GPT-5.6 Terra measured |
| LiveBench reasoning | Not measured | 90.6% | Only GPT-5.6 Terra measured |
| MMMU Pro (Vals AI) | Not measured | 86.5% | Only GPT-5.6 Terra measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 92.5% | Only GPT-5.6 Terra measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 42.9% | Only GPT-5.6 Terra measured |
| GPQA Diamond (Epoch AI run) | Not measured | 93.3% | Only GPT-5.6 Terra measured |
| ARC-AGI | Not measured | 96.5% | Only GPT-5.6 Terra measured |
| ARC-AGI-2 | Not measured | 83.9% | Only GPT-5.6 Terra measured |
| SimpleBench | Not measured | 48.9% | Only GPT-5.6 Terra measured |
| Chess puzzles | Not measured | 47.0% | Only GPT-5.6 Terra measured |
| Mystery game puzzles | Not measured | 35.0% | Only GPT-5.6 Terra measured |
| LMArena Text: hard prompts | Not measured | 1469 | Only GPT-5.6 Terra measured |
| LMArena Text: expert | Not measured | 1496 | Only GPT-5.6 Terra measured |
| LMArena Text: hard prompts in English | Not measured | 1465 | Only GPT-5.6 Terra measured |
| LMArena Vision: diagrams | Not measured | 1295 | Only GPT-5.6 Terra measured |
| LMArena Vision: homework | Not measured | 1312 | Only GPT-5.6 Terra measured |
| Kagi LLM Benchmark | Not measured | 51.3% | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 5 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| OTIS Mock AIME | Not measured | 99.7% | Only GPT-5.6 Terra measured |
| FrontierMath, tiers 1 to 3 | Not measured | 86.0% | Only GPT-5.6 Terra measured |
| FrontierMath, tier 4 | Not measured | 70.7% | Only GPT-5.6 Terra measured |
| ProofBench | Not measured | 74.0% | Only GPT-5.6 Terra measured |
| LMArena Text: math | Not measured | 1465 | Only GPT-5.6 Terra measured |
| LMArena Text: mathematical work | Not measured | 1478 | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 86.7% | Only GPT-5.6 Terra measured |
| LegalBench (Vals AI) | Not measured | 85.1% | Only GPT-5.6 Terra measured |
| CorpFin (Vals AI) | Not measured | 65.3% | Only GPT-5.6 Terra measured |
| TaxEval (Vals AI) | Not measured | 76.2% | Only GPT-5.6 Terra measured |
| MortgageTax (Vals AI) | Not measured | 67.3% | Only GPT-5.6 Terra measured |
| MedCode (Vals AI) | Not measured | 43.4% | Only GPT-5.6 Terra measured |
| SimpleQA Verified | Not measured | 43.2% | Only GPT-5.6 Terra measured |
| LMArena Text: science | Not measured | 1452 | Only GPT-5.6 Terra measured |
| LMArena Text: medicine and healthcare | Not measured | 1430 | Only GPT-5.6 Terra measured |
| LMArena Text: legal and government | Not measured | 1461 | Only GPT-5.6 Terra measured |
| LMArena Text: business and finance | Not measured | 1448 | Only GPT-5.6 Terra measured |
| FACTS Benchmark Suite | Not measured | 52.0% | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index9 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| Design Arena: websites | Not measured | 1268 | Only GPT-5.6 Terra measured |
| Design Arena: UI components | Not measured | 1256 | Only GPT-5.6 Terra measured |
| Design Arena: game dev | Not measured | 1229 | Only GPT-5.6 Terra measured |
| Design Arena: data viz | Not measured | 1240 | Only GPT-5.6 Terra measured |
| Design Arena: 3D | Not measured | 1134 | Only GPT-5.6 Terra measured |
| Design Arena: all code categories | Not measured | 1241 | Only GPT-5.6 Terra measured |
| Design Arena: SVG | Not measured | 1219 | Only GPT-5.6 Terra measured |
| Design Arena: ASCII art | Not measured | 1177 | Only GPT-5.6 Terra measured |
| LiveBench design | Not measured | 76.0% | Only GPT-5.6 Terra measured |
GeneralDirectional onlyPublic boards, shown as published, kept out of the index35 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | Leader |
|---|---|---|---|
| LiveBench data analysis | Not measured | 79.3% | Only GPT-5.6 Terra measured |
| Public Benefits Bench (Vals AI) | Not measured | 62.4% | Only GPT-5.6 Terra measured |
| SAGE (Vals AI) | Not measured | 47.0% | Only GPT-5.6 Terra measured |
| MedScribe (Vals AI) | Not measured | 82.9% | Only GPT-5.6 Terra measured |
| Vals Multimodal Index | Not measured | 65.1% | Only GPT-5.6 Terra measured |
| IFBench | Not measured | 71.2% | Only GPT-5.6 Terra measured |
| Long context reasoning (AA-LCR) | Not measured | 83.0% | Only GPT-5.6 Terra measured |
| DTBench | Not measured | 93.3% | Only GPT-5.6 Terra measured |
| LMCA | Not measured | 54.9% | Only GPT-5.6 Terra measured |
| Surface Evolver Bench | Not measured | 83.8% | Only GPT-5.6 Terra measured |
| Furniture assembly | Not measured | 54.2% | Only GPT-5.6 Terra measured |
| LMArena Text | Not measured | 1448 | Only GPT-5.6 Terra measured |
| LMArena Text: instruction following | Not measured | 1456 | Only GPT-5.6 Terra measured |
| LMArena Text: creative writing | Not measured | 1413 | Only GPT-5.6 Terra measured |
| LMArena Vision | Not measured | 1271 | Only GPT-5.6 Terra measured |
| LMArena Text: English | Not measured | 1449 | Only GPT-5.6 Terra measured |
| LMArena Text: non-English | Not measured | 1441 | Only GPT-5.6 Terra measured |
| LMArena Text: Chinese | Not measured | 1514 | Only GPT-5.6 Terra measured |
| LMArena Text: German | Not measured | 1460 | Only GPT-5.6 Terra measured |
| LMArena Text: Japanese | Not measured | 1454 | Only GPT-5.6 Terra measured |
| LMArena Text: Korean | Not measured | 1427 | Only GPT-5.6 Terra measured |
| LMArena Text: French | Not measured | 1466 | Only GPT-5.6 Terra measured |
| LMArena Text: Spanish | Not measured | 1451 | Only GPT-5.6 Terra measured |
| LMArena Text: Russian | Not measured | 1452 | Only GPT-5.6 Terra measured |
| LMArena Text: Polish | Not measured | 1446 | Only GPT-5.6 Terra measured |
| LMArena Text: multi-turn | Not measured | 1453 | Only GPT-5.6 Terra measured |
| LMArena Text: longer queries | Not measured | 1454 | Only GPT-5.6 Terra measured |
| LMArena Text: writing and language | Not measured | 1431 | Only GPT-5.6 Terra measured |
| LMArena Text: entertainment and media | Not measured | 1418 | Only GPT-5.6 Terra measured |
| LMArena Vision: English | Not measured | 1272 | Only GPT-5.6 Terra measured |
| LMArena Vision: Chinese | Not measured | 1337 | Only GPT-5.6 Terra measured |
| LMArena Vision: reading text | Not measured | 1278 | Only GPT-5.6 Terra measured |
| LMArena Vision: humor | Not measured | 1236 | Only GPT-5.6 Terra measured |
| LMArena Vision: creative writing | Not measured | 1248 | Only GPT-5.6 Terra measured |
| LMArena Document | Not measured | 1472 | Only GPT-5.6 Terra measured |
| Neither model has a result on the other 11 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Devstral 2 2512 | GPT-5.6 Terra | 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
Devstral 2 2512$0.0014Fits in one request
GPT-5.6 Terra$0.0080Fits in one request
Devstral 2 2512 costs less
Repository review
50K input + 3K output tokens
Devstral 2 2512$0.026Fits in one request
GPT-5.6 Terra$0.136Fits in one request
Devstral 2 2512 costs less
Agent loop
200K input + 10K output tokens
Devstral 2 2512$0.100Fits in one request
GPT-5.6 Terra$0.520Fits in one request
Devstral 2 2512 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 | Mistral Devstral 2 2512 | OpenAI GPT-5.6 Terra |
|---|---|---|
| Maker | Mistral AI | OpenAI |
| Weights | Open weights | No public weights listed |
| Released | December 9, 2025 | July 9, 2026 |
| Context window | 262K | 1.05M |
| Input $/Mtok | $0.40 | $2.00 |
| Output $/Mtok | $2.00 | $12.00 |
| Coding index | Not benchmarked | 63.9 |
| Independent measurements behind it | 1 | 6 |
| Cost per solved task | $1.19 | $0.42 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Devstral 2 2512 better than GPT-5.6 Terra for coding?
Only GPT-5.6 Terra has a coding score, 63.9 on the Vibe Coding Index; Devstral 2 2512 has none, so the two cannot be ranked on coding. Pick Devstral 2 2512 for output tokens 6.0x cheaper ($2.00 vs $12.00 per million).
Which is cheaper, Devstral 2 2512 or GPT-5.6 Terra?
Devstral 2 2512 costs $0.40 per million input tokens and $2.00 per million output tokens; GPT-5.6 Terra costs $2.00 and $12.00. Per solved SWE-bench Verified task, Devstral 2 2512 costs $1.19 and GPT-5.6 Terra $0.42.
Which has the bigger context window, Devstral 2 2512 or GPT-5.6 Terra?
GPT-5.6 Terra, with 1.05M tokens against 262K.
How many benchmarks is the Devstral 2 2512 vs GPT-5.6 Terra comparison based on?
5 benchmarks have a result for both models as of October 22, 2026. Devstral 2 2512's coding index rests on 1 independent benchmark and GPT-5.6 Terra's on 6. Where only one model has a result, the page says so instead of naming a winner.