Updated October 20, 2026. Rank says Spark-X2.6-4B is ahead. Price, access, and your workload can each overturn that. Public scores include evidence status and uncertainty.
Spark-X2.6-4B has the higher public point estimate, 40.14 versus 25.55. Their conditional score ranges overlap. These ranges do not establish rank confidence. 9 results are shared. Category rows resting on Estimated evidence or different benchmark sets are marked directional and do not name a winner.
NVIDIA
25.55/100
Estimated · Public rank #205
Conditional range 11.2–39.9
Recommendations appear only when a shared evidence basis or an explicit operating constraint supports the call. Secondary and unsupported use cases stay disclosed below the initial list.
No workload recommendation clears the current evidence threshold.
Use the matched evidence, workload costs, and sourced specifications below instead of treating a point score as a universal answer.
Code generation, repair, and software-engineering tasks
Not enough matched evidence
Nemotron 3.5 Lightning 30B A3B NVFP4 and Spark-X2.6-4B are scored on Estimated evidence for coding, so the reading is directional rather than like-for-like.
Tool use, computer use, and multi-step task completion
Not enough matched evidence
Nemotron 3.5 Lightning 30B A3B NVFP4 and Spark-X2.6-4B are scored on Estimated evidence for agentic, so the reading is directional rather than like-for-like.
Prompts that approach the documented context limit
No clear pick
The documented context windows are equal.
1K fresh input + 500 output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
200K cached + 20K fresh input + 10K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
50K fresh input + 3K output tokens
Not enough matched evidence
A complete comparable API-rate estimate is not available for both models.
Choose a job from the LLM Selector's task catalog to see the category row it rests on, under the same basis rules as the table below. A directional row stays directional; choosing a job never creates a winner.
The same task catalog as the LLM Selector. Each job names the evidence surface it rests on; nothing here adds a new score.
Directional only · BenchAlign v5.8
Spark-X2.6-4B has the higher coding point estimate. Conditional score ranges do not establish rank confidence.
Coding scores combine specific tasks and setups. Match the editor, harness, and effort to your workflow.
Shared results can support a head-to-head reading. Results present for only one model describe coverage, not superiority.
3 categories rest on Estimated evidence or different evidence sets. Those rows remain visible for coverage context but do not name a winner.
Only shared public evidence is shown. Sparse evidence stays a ruled list rather than being closed into a radar shape.
Too few matched category axes support a radar. The ruled list below shows only shared benchmark results; positions use each benchmark’s normalized display scale when available.
SWE MultilingualCoding
Normalized gap 28.2BrowseCompAgentic
Normalized gap 16.3SWE-bench VerifiedCoding
Normalized gap 10.0HLE w/o toolsKnowledge
Normalized gap 7.4GPQAKnowledge
Normalized gap 5.7Each row shows the public-lane category score for both models: the BenchAlign v5.8 lane for agentic, coding, and knowledge, and the weighted public rows elsewhere, with the same rank each model page reports. A row is like-for-like only when both scores rest on Supported evidence or the same weighted set. Directional and not-comparable rows remain visible, but they never receive a winner.
| Category | Nemotron 3.5 Lightning 30B A3B NVFP4 | Spark-X2.6-4B | Basis | Reading |
|---|---|---|---|---|
| Instruction following | 71.4#72/149 | 78.2#62/149 | Like-for-likeProvisional lane · 1 vs 1 weighted rows | Spark-X2.6-4B leads |
| Agentic | 15.8Estimated · #117/131 | 33.0Estimated · #89/131 | Directional onlyBenchAlign v5.8 lane · 4 vs 7 public rows | Directional only |
| Coding | 15.3Estimated · #134/152 | 30.6Estimated · #95/152 | Directional onlyBenchAlign v5.8 lane · 4 vs 4 public rows | Directional only |
| Knowledge | 29.2Estimated · #160/187 | 43.0Estimated · #109/187 | Directional onlyBenchAlign v5.8 lane · 4 vs 3 public rows | Directional only |
| Reasoning | 58.8Unranked · 1 rankable row | 48.8Unranked · 2 rankable rows | Not comparableProvisional lane · 0 vs 1 weighted rows | Not comparable |
| Multimodal | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Multilingual | Not ranked | Not ranked | Not comparableProvisional lane · 0 vs 0 weighted rows | Not comparable |
| Math | Not ranked | 76.4Unranked · 3 rankable rows | Not comparableProvisional lane · 0 vs 2 weighted rows | Not comparable |
Ranks count the models scored in each category’s lane, so the agentic, coding, and knowledge denominators (BenchAlign v5.8) differ from the provisional-lane categories. Unranked scores sit on the provisional lane’s scale but never name a winner. Like-for-like rows within 0.5 points read as a practical tie.
Three fixed token mixes turn per-token rates into comparable decisions. Each scenario states context fit and whether cached input had to fall back to the published list-input rate.
1K fresh input + 500 output tokens
Nemotron 3.5 Lightning 30B A3B NVFP4 has no comparable published API token rate. Spark-X2.6-4B has no comparable published API token rate.
50K fresh input + 3K output tokens
Nemotron 3.5 Lightning 30B A3B NVFP4 has no comparable published API token rate. Spark-X2.6-4B has no comparable published API token rate.
200K cached + 20K fresh input + 10K output tokens
Nemotron 3.5 Lightning 30B A3B NVFP4 has no comparable published API token rate. Spark-X2.6-4B has no comparable published API token rate.
Sourced differences are shown directly. Missing facts stay explicit instead of being inferred from a model name or family.
Maximum documented context; output-token limits may be lower.
Nemotron 3.5 Lightning 30B A3B NVFP4
Spark-X2.6-4B
Nemotron 3.5 Lightning 30B A3B NVFP4
Not sourced
Spark-X2.6-4B
Not sourced
A missing cached-input rate falls back to the listed input rate only in the stated workload estimate.
Nemotron 3.5 Lightning 30B A3B NVFP4
No comparable hosted API rate
NVIDIA Nemotron 3.5 Lightning model cardSpark-X2.6-4B
No comparable hosted API rate
SparkLLM Spark-X2.6-4B model cardNemotron 3.5 Lightning 30B A3B NVFP4
Not sourced
Spark-X2.6-4B
Nemotron 3.5 Lightning 30B A3B NVFP4
Not sourced
Spark-X2.6-4B
Nemotron 3.5 Lightning 30B A3B NVFP4
Not sourced
Spark-X2.6-4B
Open Weights · Hugging Face, self-hosted inference
SparkLLM Spark-X2.6-4B model cardNemotron 3.5 Lightning 30B A3B NVFP4
Reasoning
Spark-X2.6-4B
Reasoning
Nemotron 3.5 Lightning 30B A3B NVFP4
Open Weight
Spark-X2.6-4B
Open Weight
Nemotron 3.5 Lightning 30B A3B NVFP4
Open Weight
Spark-X2.6-4B
Open Weight
Nemotron 3.5 Lightning 30B A3B NVFP4
2026-08-11
Spark-X2.6-4B
2026-10-19
Spark-X2.6-4B has the higher public point estimate, 40.14 versus 25.55. Their conditional score ranges overlap. These ranges do not establish rank confidence. The page therefore keeps the decision tied to the specific documented workload.
Spark-X2.6-4B scores higher for coding on the public lane, 30.6 to 15.3. Nemotron 3.5 Lightning 30B A3B NVFP4 and Spark-X2.6-4B are scored on Estimated evidence for coding, so the reading is directional rather than like-for-like. BenchLM does not name a winner for coding from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Spark-X2.6-4B scores higher for agentic tasks on the public lane, 33 to 15.8. Nemotron 3.5 Lightning 30B A3B NVFP4 and Spark-X2.6-4B are scored on Estimated evidence for agentic tasks, so the reading is directional rather than like-for-like. BenchLM does not name a winner for agentic tasks from a directional reading; read the shared benchmark rows directly and test the models on the same task set.
Both models do not have comparable published API token rates, so this page does not name a universal price winner.
Both models list the same context window, 1M.
Estimates at 50,000 req/day · 1000 tokens/req average.
The full public result ledger is available for audit without forcing a wide desktop table onto a phone.
Terminal-Bench 2.1
Not directly comparable
PinchBench
Not directly comparable
BrowseComp
Spark-X2.6-4B leads this result
τ³-bench results
Spark-X2.6-4B leads this result
BFCL v4
Not directly comparable
τ²-bench results
Not directly comparable
MCP Atlas
Not directly comparable
HLE w/ tools
Not directly comparable
Claw-Eval
Not directly comparable
SWE-bench Verified
Spark-X2.6-4B leads this result
SWE Multilingual
Spark-X2.6-4B leads this result
Terminal-Bench 2.1
Not directly comparable
SciCode
Spark-X2.6-4B leads this result
SWE-bench Pro
Not directly comparable
LongBench v2
Not directly comparable
GPQA
Nemotron 3.5 Lightning 30B A3B NVFP4 leads this result
GPQA-D
Spark-X2.6-4B leads this result
HLE w/o tools
Spark-X2.6-4B leads this result
MMLU-Pro
Not directly comparable
One weekly email when material rank, price, or benchmark evidence changes make this matchup worth revisiting.
Read a sample issueJoin 5,600+ readers.
Last updated October 20, 2026