Codestral 2508 vs Command A
Codestral 2508 costs 11.1x less per output token than Command A ($0.90 vs $10.00 per million). Updated October 21, 2026; 1 benchmarks have a result for both.
Codestral 2508 or Command A: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Codestral 2508 for output tokens 11.1x cheaper ($0.90 vs $10.00 per million).
- Shared results
- 1
- Codestral 2508 only
- 7
- Command A only
- 61
- Like-for-like categories
- 1 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
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
No call
Only Command A has a AA Intelligence Index result, 13.1.
Directional only
Agentic work
AA Agentic Index
No call
Only Command A has a AA Agentic Index result, 1.0.
Directional only
Reasoning
GPQA Diamond
No call
Only Command A has a GPQA Diamond result, 48.5%.
Directional only
Math
AIME
No call
Only Command A has a AIME result, 13.3%.
Directional only
Knowledge
MMLU Pro
No call
Only Command A has a MMLU Pro result, 69.2%.
Directional only
Web design
Design Arena: websites
No call
Only Codestral 2508 has a Design Arena: websites result, 1013.
Directional only
Long documents
Listed context window
No call
Both take 256K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Codestral 2508
Codestral 2508 $0.0008, Command A $0.0075 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Command A has the number, $2.69.
Directional only
At a glance
- Coding index
- Codestral 2508 not benchmarked, Command A not benchmarked.
- Widest category
- Reasoning, 3.7 points apart on average over 1 shared benchmark.
- Widest single gap
- Kagi LLM Benchmark: Codestral 2508 32.5% against Command A 28.8%, 3.7 points apart on a 0 to 100 scale.
- Output price
- Codestral 2508 $0.90 against Command A $10.00 per million output tokens, 11.1x apart.
- Context window
- Both take 256K tokens.
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 | Codestral 2508 | Bars | Command A | Ties |
|---|---|---|---|---|
| Reasoning | 1 | 0 | ||
| All shared boards | 1 | 0 |
CodingDirectional onlySix of these boards feed the Vibe Coding Index10 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| AA Coding Index | Not measured | 27.8 | Only Command A measured |
| LiveCodeBench (Vals AI) | Not measured | 35.1% | Only Command A measured |
| SWE-bench Verified (Vals AI) | Not measured | 7.8% | Only Command A measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 17.6% | Only Command A measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 28.7% | Only Command A measured |
| SciCode | Not measured | 38.5% | Only Command A measured |
| Terminal-Bench Hard (Artificial Analysis run) | Not measured | 25.0% | Only Command A measured |
| ALE-Bench | 137.8 | Not measured | Only Codestral 2508 measured |
| LMArena Text: coding | Not measured | 1330 | Only Command A measured |
| LMArena Text: software and IT | Not measured | 1336 | Only Command A measured |
| Neither model has a result on the other 29 boards here | |||
IntelligenceDirectional onlyPublic boards, shown as published, kept out of the index1 board, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| AA Intelligence Index | Not measured | 13.1 | Only Command A measured |
| Neither model has a result on the other 1 board here | |||
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 1.0 | Only Command A measured |
| Finance Agent (Vals AI) | Not measured | 9.0% | Only Command A measured |
| τ²-Bench Telecom | Not measured | 80.7% | Only Command A measured |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 46.5% | Only Command A measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 48.5% | Only Command A measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 76.1% | Only Command A measured |
| Humanity's Last Exam (Artificial Analysis run) | Not measured | 12.0% | Only Command A measured |
| LMArena Text: hard prompts | Not measured | 1326 | Only Command A measured |
| LMArena Text: expert | Not measured | 1296 | Only Command A measured |
| LMArena Text: hard prompts in English | Not measured | 1335 | Only Command A measured |
| Kagi LLM Benchmark | 32.5% | 28.8% | Codestral 2508 |
| Neither model has a result on the other 15 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| AIME | Not measured | 13.3% | Only Command A measured |
| MATH 500 (Vals AI) | Not measured | 76.2% | Only Command A measured |
| MGSM (Vals AI) | Not measured | 85.7% | Only Command A measured |
| AA Math Index | Not measured | 13.0 | Only Command A measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 13.0% | Only Command A measured |
| LMArena Text: math | Not measured | 1300 | Only Command A measured |
| LMArena Text: mathematical work | Not measured | 1312 | Only Command A measured |
| Neither model has a result on the other 6 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 69.2% | Only Command A measured |
| LegalBench (Vals AI) | Not measured | 79.7% | Only Command A measured |
| CaseLaw (Vals AI) | Not measured | 64.5% | Only Command A measured |
| CorpFin (Vals AI) | Not measured | 46.0% | Only Command A measured |
| TaxEval (Vals AI) | Not measured | 61.4% | Only Command A measured |
| MedQA (Vals AI) | Not measured | 80.5% | Only Command A measured |
| MedCode (Vals AI) | Not measured | 19.7% | Only Command A measured |
| MMLU Pro (Artificial Analysis run) | Not measured | 71.2% | Only Command A measured |
| LMArena Text: science | Not measured | 1339 | Only Command A measured |
| LMArena Text: medicine and healthcare | Not measured | 1334 | Only Command A measured |
| LMArena Text: legal and government | Not measured | 1348 | Only Command A measured |
| LMArena Text: business and finance | Not measured | 1315 | Only Command A measured |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| Design Arena: websites | 1013 | Not measured | Only Codestral 2508 measured |
| Design Arena: UI components | 1017 | Not measured | Only Codestral 2508 measured |
| Design Arena: game dev | 968 | Not measured | Only Codestral 2508 measured |
| Design Arena: data viz | 1017 | Not measured | Only Codestral 2508 measured |
| Design Arena: 3D | 1020 | Not measured | Only Codestral 2508 measured |
| Design Arena: all code categories | 1007 | Not measured | Only Codestral 2508 measured |
| Neither model has a result on the other 3 boards here | |||
GeneralDirectional onlyPublic boards, shown as published, kept out of the index22 boards, none shared
| Benchmark | Codestral 2508 | Command A | Leader |
|---|---|---|---|
| MedScribe (Vals AI) | Not measured | 55.7% | Only Command A measured |
| IFBench | Not measured | 73.9% | Only Command A measured |
| Long context reasoning (AA-LCR) | Not measured | 52.7% | Only Command A measured |
| DTBench | Not measured | 61.3% | Only Command A measured |
| LMCA | Not measured | 10.3% | Only Command A measured |
| LMArena Text | Not measured | 1331 | Only Command A measured |
| LMArena Text: instruction following | Not measured | 1309 | Only Command A measured |
| LMArena Text: creative writing | Not measured | 1319 | Only Command A measured |
| LMArena Text: English | Not measured | 1344 | Only Command A measured |
| LMArena Text: non-English | Not measured | 1313 | Only Command A measured |
| LMArena Text: Chinese | Not measured | 1327 | Only Command A measured |
| LMArena Text: German | Not measured | 1342 | Only Command A measured |
| LMArena Text: Japanese | Not measured | 1285 | Only Command A measured |
| LMArena Text: Korean | Not measured | 1285 | Only Command A measured |
| LMArena Text: French | Not measured | 1351 | Only Command A measured |
| LMArena Text: Spanish | Not measured | 1347 | Only Command A measured |
| LMArena Text: Russian | Not measured | 1314 | Only Command A measured |
| LMArena Text: Polish | Not measured | 1324 | Only Command A measured |
| LMArena Text: multi-turn | Not measured | 1338 | Only Command A measured |
| LMArena Text: longer queries | Not measured | 1334 | Only Command A measured |
| LMArena Text: writing and language | Not measured | 1328 | Only Command A measured |
| LMArena Text: entertainment and media | Not measured | 1306 | Only Command A measured |
| Neither model has a result on the other 24 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Codestral 2508 | Command A | 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
Codestral 2508 costs less
Repository review
50K input + 3K output tokens
Codestral 2508 costs less
Agent loop
200K input + 10K output tokens
Codestral 2508 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 Codestral 2508 | Cohere Command A |
|---|---|---|
| Maker | Mistral AI | Cohere |
| Weights | No public weights listed | Open weights |
| Released | August 1, 2025 | March 13, 2025 |
| Context window | 256K | 256K |
| Input $/Mtok | $0.30 | $2.50 |
| Output $/Mtok | $0.90 | $10.00 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 0 | 1 |
| Cost per solved task | Not computable | $2.69 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Codestral 2508 better than Command A for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Codestral 2508 for output tokens 11.1x cheaper ($0.90 vs $10.00 per million).
Which is cheaper, Codestral 2508 or Command A?
Codestral 2508 costs $0.30 per million input tokens and $0.90 per million output tokens; Command A costs $2.50 and $10.00. Cost per solved task is known for Command A only.
Which has the bigger context window, Codestral 2508 or Command A?
Neither: both take 256K tokens.
How many benchmarks is the Codestral 2508 vs Command A comparison based on?
1 benchmarks have a result for both models as of October 21, 2026. Codestral 2508's coding index rests on 0 independent benchmarks and Command A's on 1. Where only one model has a result, the page says so instead of naming a winner.